{"type": "FeatureCollection", "features": [{"id": "10.1002/adma.202006054", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:14:19Z", "type": "Journal Article", "created": "2021-02-22", "title": "Infrared Metasurface Augmented by Deep Learning for Monitoring Dynamics between All Major Classes of Biomolecules", "description": "Abstract

Insights into the fascinating molecular world of biological processes are crucial for understanding diseases, developing diagnostics, and effective therapeutics. These processes are complex as they involve interactions between four major classes of biomolecules, i.e., proteins, nucleic acids, carbohydrates, and lipids, which makes it important to be able to discriminate between all these different biomolecular species. In this work, a deep learning\uffe2\uff80\uff90augmented, chemically\uffe2\uff80\uff90specific nanoplasmonic technique that enables such a feat in a label\uffe2\uff80\uff90free manner to not disrupt native processes is presented. The method uses a highly sensitive multiresonant plasmonic metasurface in a microfluidic device, which enhances infrared absorption across a broadband mid\uffe2\uff80\uff90IR spectrum and in water, despite its strongly overlapping absorption bands. The real\uffe2\uff80\uff90time format of the optofluidic method enables the collection of a vast amount of spectrotemporal data, which allows the construction of a deep neural network to discriminate accurately between all major classes of biomolecules. The capabilities of the new method are demonstrated by monitoring of a multistep bioassay containing sucrose\uffe2\uff80\uff90 and nucleotides\uffe2\uff80\uff90loaded liposomes interacting with a small, lipid membrane\uffe2\uff80\uff90perforating peptide. It is envisioned that the presented technology will impact the fields of biology, bioanalytics, and pharmacology from fundamental research and disease diagnostics to drug development.2 = 0.73), followed by Cubist (validation R2 = 0.66) and memorybased learning (validation R2 = 0.65). After LULC stratification, Cubist model achieved the best prediction (validation R2 = 0.76), improving the value of ratio of performance to interquartile distance by 0.11 compared to the global CNN model. Areas with high SOC values were mainly located in the city center. Stratification by LULC class is a promising strategy for addressing the impact of the soil-landscape diversity and complexity on vis\u2013NIR spectral estimation of SOC in urban soil spectral library.", "keywords": ["Urban soil", "Stratified modeling", "13. Climate action", "Soil organic carbon", "11. Sustainability", "0401 agriculture", " forestry", " and fisheries", "Deep learning", "04 agricultural and veterinary sciences", "15. Life on land", "Soil spectral library"]}, "links": [{"href": "https://doi.org/10.1016/j.geoderma.2022.116102"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Geoderma", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.1016/j.geoderma.2022.116102", "name": "item", "description": "10.1016/j.geoderma.2022.116102", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.1016/j.geoderma.2022.116102"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2022-11-01T00:00:00Z"}}, {"id": "10.1038/s41586-023-06042-3", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:17:51Z", "type": "Journal Article", "created": "2023-05-24", "title": "Microbial carbon use efficiency promotes global soil carbon storage", "description": "Abstract

Soils store more carbon than other terrestrial ecosystems1,2. How soil organic carbon (SOC) forms and persists remains uncertain1,3, which makes it challenging to understand how it will respond to climatic change3,4. It has been suggested that soil microorganisms play an important role in SOC formation, preservation and loss5\uffe2\uff80\uff937. Although microorganisms affect the accumulation and loss of soil organic matter through many pathways4,6,8\uffe2\uff80\uff9311, microbial carbon use efficiency (CUE) is an integrative metric that can capture the balance of these processes12,13. Although CUE has the potential to act as a predictor of variation in SOC storage, the role of CUE in SOC persistence remains unresolved7,14,15. Here we examine the relationship between CUE and the preservation of SOC, and interactions with climate, vegetation and edaphic properties, using a combination of global-scale datasets, a microbial-process explicit model, data assimilation, deep learning and meta-analysis. We find that CUE is at least four times as important as other evaluated factors, such as carbon input, decomposition or vertical transport, in determining SOC storage and its spatial variation across the globe. In addition, CUE shows a positive correlation with SOC content. Our findings point to microbial CUE as a major determinant of global SOC storage. Understanding the microbial processes underlying CUE and their environmental dependence may help the prediction of SOC feedback to a changing climate.Tomato is an important commercial product which is perishable by nature and highly susceptible to fungal incidence once it is harvested. Not all tomatoes are equally vulnerable to pathogenic fungi, and an early detection of the vulnerable ones can help in taking timely preventive actions, ranging from isolating tomato batches to adjusting storage conditions, but also in making right business decisions like dynamic pricing based on quality or better shelf life estimate. More importantly, early detection of vulnerable produce can help in taking timely actions to minimize potential post-harvest losses. This paper investigates Near-infrared (NIR) hyperspectral imaging (1000\uffe2\uff80\uff931700\uffc2\uffa0nm) and machine learning to build models to automatically predict the susceptibility of sepals of recently harvested tomatoes to future fungal infections. Hyperspectral images of newly harvested tomatoes (cultivar Brioso) from 5 different growers were acquired before the onset of any visible fungal infection. After imaging, the tomatoes were placed under controlled conditions suited for fungal germination and growth for a 4-day period, and then imaged using normal color cameras. All sepals in the color images were ranked for fungal severity using crowdsourcing, and the final severity of each sepal was fused using principal component analysis. A novel hyperspectral data processing pipeline is presented which was used to automatically segment the tomato sepals from spectral images with multiple tomatoes connected via a truss. The key modelling question addressed in this research is whether there is a correlation between the hyperspectral data captured at harvest and the fungal infection observed 4 days later. Using 10-fold and group k-fold cross-validation, XG-Boost and Random Forest based regression models were trained on the features derived from the hyperspectral data corresponding to each sepal in the training set and tested on hold out test set. The best model found a Pearson correlation of 0.837, showing that there is strong linear correlation between the NIR spectra and the future fungal severity of the sepal. The sepal specific predictions were aggregated to predict the susceptibility of individual tomatoes, and a correlation of 0.92 was found. Besides modelling, focus is also on model interpretation, particularly to understand which spectral features are most relevant to model prediction. Two approaches to model interpretation were explored, feature importance and SHAP (SHapley Additive exPlanations), resulting in similar conclusions that the NIR range between 1390\uffe2\uff80\uff931420\uffc2\uffa0nm contributes most to the model\uffe2\uff80\uff99s final decision.The scale of root quantification in research is often limited by the time required for sampling, measurement and processing samples. Recent developments in Convolutional Neural Networks (CNN) have made faster and more accurate plant image analysis possible which may significantly reduce the time required for root measurement, but challenges remain in making these methods accessible to researchers without an in-depth knowledge of Machine Learning. We analyzed root images acquired from three destructive root samplings using the RootPainter CNN-software that features an interface for corrective annotation for easier use. Root scans with and without non-root debris were used to test if training a model, i.e., learning from labeled examples, can effectively exclude the debris by comparing the end-results with measurements from clean images. Root images acquired from soil profile walls and the cross-section of soil cores were also used for training and the derived measurements were compared with manual measurements. After 200 minutes of training on each dataset, significant relationships between manual measurements and RootPainter-derived data were noted for monolith (R2=0.99), profile wall (R2=0.76) and core-break (R2=0.57). The rooting density derived from images with debris was not significantly different from that derived from clean images after processing with RootPainter. Rooting density was also successfully calculated from both profile wall and soil core images, and in each case the gradient of root density with depth was not significantly different from manual counts. Our results demonstrate that the proposed approach using CNN can lead to substantial reductions in root sample processing workloads, increasing the potential scale of future root investigations.

", "keywords": ["0301 basic medicine", "root phenotyping", "profile wall", "root washing", "segmentation", "deep learning", "Convolutional neural network", "04 agricultural and veterinary sciences", "15. Life on land", "Soil", "03 medical and health sciences", "core-break", "monolith", "soil coring", "Image Processing", " Computer-Assisted", "0401 agriculture", " forestry", " and fisheries", "Neural Networks", " Computer", "Software"]}, "links": [{"href": "http://academic.oup.com/jxb/article-pdf/72/13/4680/38807872/erab174.pdf"}, {"href": "https://doi.org/10.1093/jxb/erab174"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Journal%20of%20Experimental%20Botany", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.1093/jxb/erab174", "name": "item", "description": "10.1093/jxb/erab174", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.1093/jxb/erab174"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2020-12-02T00:00:00Z"}}, {"id": "10.1109/access.2023.3339884", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:18:35Z", "type": "Journal Article", "created": "2023-12-05", "title": "Classifying the Vermicompost Production Stages Using Thermal Camera Data", "description": "The procedure of processing the vermicompost production includes several stages, where the vermicompost material has different temperatures during these different stages. Thermal sensors play a key role in numerous fields, such as medical and agricultural applications. Thermal cameras can produce a thermal image or an array of values representing the array of sensory data. i.e., an array of temperatures. In this study, we proposed the first thermal imagery dataset of the vermicompost production process. The contributions of this work are two-fold using the proposed dataset. First, we framed the process of predicting the vermicompost production process as a classification problem. Second, we compared classifying the different stages of the process of vermicompost production based on two different input types, namely, thermal images and an array of temperatures. In other words, the classifier will be fed with an input (an image or an array of temperatures), and then the classifier will predict the vermicompost production stage. In this context, we utilized several machine and deep learning models as classifiers. For the utilized dataset, the study has been conducted on a set of images collected during the vermicompost production procedure which was collected every 14 days over 42 consecutive days, i.e., four classes. We proposed running a series of experiments to determine which input type yields better classification accuracy. The obtained results show that using thermal images for the sake of classifying the vermicompost production stages achieved higher accuracy, about 92%, in comparison to using the sensor array data, about 60%.", "keywords": ["machine learning", "SENet", "deep learning", "Electrical engineering. Electronics. Nuclear engineering", "sensor array", "Classification", "ResNet", "TK1-9971"]}, "links": [{"href": "https://doi.org/10.1109/access.2023.3339884"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/IEEE%20Access", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.1109/access.2023.3339884", "name": "item", "description": "10.1109/access.2023.3339884", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.1109/access.2023.3339884"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2023-01-01T00:00:00Z"}}, {"id": "10.1109/access.2025.3569213", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:18:35Z", "type": "Journal Article", "created": "2025-05-12", "title": "An Efficient Encoding Spectral Information in Hyperspectral Images for Transfer Learning of Mask R-CNN for Instance Segmentation of Tomato Sepals", "description": "The most vulnerable parts of tomatoes are the tips of the sepals, which are the primary entry points for fungal spores. Their precise segmentation within hyperspectral images (HSIs) plays a pivotal role in the development of automated and non-destructive systems for assessing tomatoes’ sensitivity to fungal infections. This research addresses the critical need for encoding spectral information in hyperspectral imaging to enhance the efficiency of such automated systems. We investigate four different techniques: Principal Component Analysis (PCA), Independent Component Analysis (ICA), Probabilistic Principal Component Analysis (PPCA), and Non-Negative Matrix Factorization (NMF), to perform transfer learning for tomato sepal instance segmentation using models previously trained on RGB images. A comparative analysis of three Mask Region-based Convolutional Neural Network (Mask R-CNN) backbone models is conducted: the Faster R-CNN, Deformable ConvNet, and Feature Pyramid Network (FPN) on spectral-encoded HSIs of the Brioso tomato variety. The Mask R-CNN with FPN, integrated with the NMF technique achieved the highest level of accuracy, yielding a Mean Average Precision (mAP) of 94.05%. Furthermore, on the second dataset, which included an additional three tomato varieties: Capricia, Provine, and Sao Paolo, the same model achieved mAP score of 86.42% across all tomato varieties with only a single false positive detection. Additionally, we incorporated a custom convolutional layer initialized it with estimated NMF coefficients, and achieved a mAP score of 87.40%. This demonstrates the potential of integrating spectral information encoding with trained deep learning-based instance segmentation models to enable robust and accurate transfer learning for automated agricultural food quality assessments.", "keywords": ["instance segmentation", "deep learning", "Electrical engineering. Electronics. Nuclear engineering", "transfer learning", "tomato", "encoding spectral information in HSI", "Hyperspectral imaging (HSI)", "TK1-9971"]}, "links": [{"href": "https://doi.org/10.1109/access.2025.3569213"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/IEEE%20Access", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.1109/access.2025.3569213", "name": "item", "description": "10.1109/access.2025.3569213", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.1109/access.2025.3569213"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2025-01-01T00:00:00Z"}}, {"id": "10.1111/nph.18387", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:19:06Z", "type": "Journal Article", "created": "2020-04-18", "title": "RootPainter: deep learning segmentation of biological images with corrective annotation", "description": "

We present RootPainter, a GUI-based software tool for the rapid training of deep neural networks for use in biological image analysis. RootPainter facilitates both fully-automatic and semi-automatic image segmentation. We investigate the effectiveness of RootPainter using three plant image datasets, evaluating its potential for root length extraction from chicory roots in soil, biopore counting and root nodule counting from scanned roots. We also use RootPainter to compare dense annotations to corrective ones which are added during the training based on the weaknesses of the current model.

", "keywords": ["Buildings and machinery", "0301 basic medicine", "phenotyping", "root nodule", "biopore", "interactive machine learning", "Research", "segmentation", "deep learning", "rhizotron", "Breeding and genetics", "Machine Learning", "Soil", "03 medical and health sciences", "Deep Learning", "GUI", "Farm nutrient management", "Image Processing", " Computer-Assisted", "Neural Networks", " Computer"]}, "links": [{"href": "https://www.biorxiv.org/content/10.1101/2020.04.16.044461v1.full.pdf"}, {"href": "https://nph.onlinelibrary.wiley.com/doi/pdf/10.1111/nph.18387"}, {"href": "https://doi.org/10.1111/nph.18387"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/New%20Phytologist", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.1111/nph.18387", "name": "item", "description": "10.1111/nph.18387", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.1111/nph.18387"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2020-04-18T00:00:00Z"}}, {"id": "10.13140/rg.2.2.35593.04962", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:19:27Z", "type": "Report", "title": "Testing a SAR-based ship classifier with different loss functions", "description": "This study investigated the influence of six different loss functions on Synthetic Aperture Radar (SAR) ship classification accuracy across two datasets. Kullback-Leibler Divergence Loss emerged with the highest average accuracy (69.5%), followed by L1 Loss (69.12%) and Focal Loss(68.4%). Interestingly, L1 and Focal Loss exhibited contrasting performance across datasets, suggesting potential data-specific suitability for certain functions. These findings highlight the importance of considering data characteristics and task requirements when selecting loss functions to optimize SAR ship classification performance.", "keywords": ["Loss Functions", " Deep Learning", " Ship Classification", " SAR"], "contacts": [{"organization": "Awais Ch. M., Reggiannini M., Moroni D.,", "roles": ["creator"]}]}, "links": [{"href": "https://iris.cnr.it/bitstream/20.500.14243/517581/1/TestingSARbasedShipClassifierwithDifferentLossFunctions.pdf"}, {"href": "https://doi.org/10.13140/rg.2.2.35593.04962"}, {"rel": "self", "type": "application/geo+json", "title": "10.13140/rg.2.2.35593.04962", "name": "item", "description": "10.13140/rg.2.2.35593.04962", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.13140/rg.2.2.35593.04962"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2024-01-01T00:00:00Z"}}, {"id": "10.3390/app142210693", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:20:56Z", "type": "Journal Article", "created": "2024-11-19", "title": "Corn Plant In-Row Distance Analysis Based on Unmanned Aerial Vehicle Imagery and Row-Unit Dynamics", "description": "

Uniform spatial distribution of plants is crucial in arable crops. Seeding quality is affected by numerous parameters, including the working speed and vibrations of the seeder. Therefore, investigating effective and rapid methods to evaluate seeding quality and the parameters affecting the seeders\u2019 performance is of high importance. With the latest advancements in unmanned aerial vehicle (UAV) technology, the potential for acquiring accurate agricultural data has significantly increased, making UAVs an ideal tool for scouting applications in agricultural systems. This study investigates the effectiveness of utilizing different plant recognition algorithms applied to UAV-derived images for evaluating seeder performance based on detected plant spacings. Additionally, it examines the impact of seeding unit vibrations on seeding quality by analyzing accelerometer data installed on the seeder. For the image analysis, three plant recognition approaches were tested: an unsupervised segmentation method based on the Visible Atmospherically Resistant Index (VARI), template matching (TM), and a deep learning model called Mask R-CNN. The Mask R-CNN model demonstrated the highest recognition reliability at 96.7%, excelling in detecting seeding errors such as misses and doubles, as well as in evaluating the quality of feed index and precision when compared to ground-truth data. Although the VARI-based unsupervised method and TM outperformed Mask R-CNN in recognizing double spacings, overall, the Mask R-CNN was the most promising. Vibration analysis indicated that the seeder\u2019s working speed significantly affected seeding quality. These findings suggest areas for potential improvements in machine technology to improve sowing operations.

", "keywords": ["0106 biological sciences", "Technology", "QH301-705.5", "T", "Physics", "QC1-999", "UAV imagery", "deep learning", "04 agricultural and veterinary sciences", "Engineering (General). Civil engineering (General)", "01 natural sciences", "FFT", "Chemistry", "seeding quality", "0401 agriculture", " forestry", " and fisheries", "TA1-2040", "Biology (General)", "QD1-999", "vibration analysis"]}, "links": [{"href": "https://www.mdpi.com/2076-3417/14/22/10693/pdf"}, {"href": "https://doi.org/10.3390/app142210693"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Applied%20Sciences", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/app142210693", "name": "item", "description": "10.3390/app142210693", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/app142210693"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2024-11-19T00:00:00Z"}}, {"id": "10.3390/agriculture15020164", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:20:53Z", "type": "Journal Article", "created": "2025-01-13", "title": "Comparative Evaluation of AI-Based Multi-Spectral Imaging and PCR-Based Assays for Early Detection of Botrytis cinerea Infection on Pepper Plants", "description": "

Pepper production is a critical component of the global agricultural economy, with exports reaching a remarkable $6.9B in 2023. This underscores the crop\u2019s importance as a major economic driver of export revenue for producing nations. Botrytis cinerea, the causative agent of gray mold, significantly impacts crops like fruits and vegetables, including peppers. Early detection of this pathogen is crucial for a reduction in fungicide reliance and economic loss prevention. Traditionally, visual inspection has been a primary method for detection. However, symptoms often appear after the pathogen has begun to spread. This study employs the Deep Learning algorithm YOLO for single-class segmentation on plant images to extract spatial details of pepper leaves. The dataset included hyperspectral images at discrete wavelengths (460 nm, 540 nm, 640 nm, 775 nm, and 875 nm) from derived vegetation indices (CVI, GNDVI, NDVI, NPCI, and PSRI) and from RGB. At an Intersection over Union with a 0.5 threshold, the Mean Average Precision (mAP50) achieved by the leaf-segmentation solution YOLOv11-Small was 86.4%. The extracted leaf segments were processed by multiple Transformer models, each yielding a descriptor. These descriptors were combined in ensemble and classified into three distinct classes using a K-nearest neighbor, a Long Short-Term Memory (LSTM), and a ResNet solution. The Transformer models that comprised the best ensemble classifier were as follows: the Swin-L (P:4 \u00d7 4\u2013W:12 \u00d7 12), the ViT-L (P:16 \u00d7 16), the VOLO (D:5), and the XCIT-L (L:24\u2013P:16 \u00d7 16), with the LSTM-based classification solution on the RGB, CVI, GNDVI, NDVI, and PSRI image sets. The classifier achieved an overall accuracy of 87.42% with an F1-Score of 81.13%. The per-class F1-Scores for the three classes were 85.25%, 66.67%, and 78.26%, respectively. Moreover, for B. cinerea detection during the initial as well as quiescent stages of infection prior to symptom development, qPCR-based methods (RT-qPCR) were used for quantification of in planta fungal biomass and integrated with the findings from the AI approach to offer a comprehensive strategy. The study demonstrates early and accurate detection of B. cinerea on pepper plants by combining segmentation techniques with Transformer model descriptors, ensembled for classification. This approach marks a significant step forward in the detection and management of crop diseases, highlighting the potential to integrate such methods into in situ systems like mobile apps or robots.

", "keywords": ["vision transformers", "Agriculture (General)", "segmentation", "deep learning", "Botrytis cinerea", "descriptor classification", "image classification", "S1-972"]}, "links": [{"href": "https://www.mdpi.com/2077-0472/15/2/164/pdf"}, {"href": "https://doi.org/10.3390/agriculture15020164"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Agriculture", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/agriculture15020164", "name": "item", "description": "10.3390/agriculture15020164", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/agriculture15020164"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2025-01-13T00:00:00Z"}}, {"id": "10.3390/agriengineering7020029", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:20:53Z", "type": "Journal Article", "created": "2025-01-27", "title": "AI-Driven Insect Detection, Real-Time Monitoring, and Population Forecasting in Greenhouses", "description": "

Insecticide use in agriculture has significantly increased over the past decades, reaching 774 thousand metric tons in 2022. This widespread reliance on chemical insecticides has substantial economic, environmental, and human health consequences, highlighting the urgent need for sustainable pest management strategies. Early detection, insect monitoring, and population forecasting through Artificial Intelligence (AI)-based methods, can enable swift responsiveness, allowing for reduced but more effective insecticide use, mitigating traditional labor-intensive and error prone solutions. The main challenge is creating AI models that perform with speed and accuracy, enabling immediate farmer action. This study highlights the innovating potential of such an approach, focusing on the detection and prediction of black aphids under state-of-the-art Deep Learning (DL) models. A dataset of 220 sticky paper images was captured. The detection system employs a YOLOv10 DL model that achieved an accuracy of 89.1% (mAP50). For insect population prediction, random forests, gradient boosting, LSTM, and the ARIMA, ARIMAX, and SARIMAX models were evaluated. The ARIMAX model performed best with a Mean Square Error (MSE) of 75.61, corresponding to an average deviation of 8.61 insects per day between predicted and actual insect counts. For the visualization of the detection results, the DL model was embedded to a mobile application. This holistic approach supports early intervention strategies and sustainable pest management while offering a scalable solution for smart-agriculture environments.

", "keywords": ["machine learning", "Agriculture (General)", "insect detection", "deep learning", "black aphids", "mobile application", "TA1-2040", "Engineering (General). Civil engineering (General)", "insect population prediction", "S1-972"]}, "links": [{"href": "https://www.mdpi.com/2624-7402/7/2/29/pdf"}, {"href": "https://doi.org/10.3390/agriengineering7020029"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/AgriEngineering", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/agriengineering7020029", "name": "item", "description": "10.3390/agriengineering7020029", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/agriengineering7020029"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2025-01-27T00:00:00Z"}}, {"id": "10.3390/drones7030179", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:20:57Z", "type": "Journal Article", "created": "2023-03-07", "title": "Animal Detection and Counting from UAV Images Using Convolutional Neural Networks", "description": "

In the last decade, small unmanned aerial vehicles (UAVs/drones) have become increasingly popular in the airborne observation of large areas for many purposes, such as the monitoring of agricultural areas, the tracking of wild animals in their natural habitats, and the counting of livestock. Coupled with deep learning, they allow for automatic image processing and recognition. The aim of this work was to detect and count the deer population in northwestern Serbia from such images using deep neural networks, a tedious process that otherwise requires a lot of time and effort. In this paper, we present and compare the performance of several state-of-the-art network architectures, trained on a manually annotated set of images, and use it to predict the presence of objects in the rest of the dataset. We implemented three versions of the You Only Look Once (YOLO) architecture and a Single Shot Multibox Detector (SSD) to detect deer in a dense forest environment and measured their performance based on mean average precision (mAP), precision, recall, and F1 score. Moreover, we also evaluated the models based on their real-time performance. The results showed that the selected models were able to detect deer with a mean average precision of up to 70.45% and a confidence score of up to a 99%. The highest precision was achieved by the fourth version of YOLO with 86%, as well as the highest recall value of 75%. Its compressed version achieved slightly lower results, with 83% mAP in its best case, but it demonstrated four times better real-time performance. The counting function was applied on the best-performing models, providing us with the exact distribution of deer over all images. Yolov4 obtained an error of 8.3% in counting, while Yolov4-tiny mistook 12 deer, which accounted for an error of 7.1%.

", "keywords": ["convolutional neural networks; deer; animal counting; deep learning; YOLO; SSD", "convolutional neural networks", "deer", "deep learning", "YOLO", "TL1-4050", "animal counting", "SSD", "Motor vehicles. Aeronautics. Astronautics"]}, "links": [{"href": "http://www.mdpi.com/2504-446X/7/3/179/pdf"}, {"href": "https://www.mdpi.com/2504-446X/7/3/179/pdf"}, {"href": "https://doi.org/10.3390/drones7030179"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Drones", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/drones7030179", "name": "item", "description": "10.3390/drones7030179", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/drones7030179"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2023-03-06T00:00:00Z"}}, {"id": "10.3390/ijms24076573", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:21:00Z", "type": "Journal Article", "created": "2023-04-03", "title": "A Systematic Review of Deep Learning Methodologies Used in the Drug Discovery Process with Emphasis on In Vivo Validation", "description": "

The discovery and development of new drugs are extremely long and costly processes. Recent progress in artificial intelligence has made a positive impact on the drug development pipeline. Numerous challenges have been addressed with the growing exploitation of drug-related data and the advancement of deep learning technology. Several model frameworks have been proposed to enhance the performance of deep learning algorithms in molecular design. However, only a few have had an immediate impact on drug development since computational results may not be confirmed experimentally. This systematic review aims to summarize the different deep learning architectures used in the drug discovery process and are validated with further in vivo experiments. For each presented study, the proposed molecule or peptide that has been generated or identified by the deep learning model has been biologically evaluated in animal models. These state-of-the-art studies highlight that even if artificial intelligence in drug discovery is still in its infancy, it has great potential to accelerate the drug discovery cycle, reduce the required costs, and contribute to the integration of the 3R (Replacement, Reduction, Refinement) principles. Out of all the reviewed scientific articles, seven algorithms were identified: recurrent neural networks, specifically, long short-term memory (LSTM-RNNs), Autoencoders (AEs) and their Wasserstein Autoencoders (WAEs) and Variational Autoencoders (VAEs) variants; Convolutional Neural Networks (CNNs); Direct Message Passing Neural Networks (D-MPNNs); and Multitask Deep Neural Networks (MTDNNs). LSTM-RNNs were the most used architectures with molecules or peptide sequences as inputs.

", "keywords": ["Deep Learning", "Artificial Intelligence", "Drug Discovery", "Review", "Neural Networks", " Computer", "drug discovery; drug design; artificial intelligence; machine learning; deep learning; biological evaluation; animal model; in vivo", "Algorithms", "3. Good health"]}, "links": [{"href": "https://www.mdpi.com/1422-0067/24/7/6573/pdf"}, {"href": "https://doi.org/10.3390/ijms24076573"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/International%20Journal%20of%20Molecular%20Sciences", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/ijms24076573", "name": "item", "description": "10.3390/ijms24076573", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/ijms24076573"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2023-03-31T00:00:00Z"}}, {"id": "10.3390/rs12040638", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:21:07Z", "type": "Journal Article", "created": "2020-02-20", "title": "Historical Aerial Surveys Map Long-Term Changes of Forest Cover and Structure in the Central Congo Basin", "description": "

Given the impact of tropical forest disturbances on atmospheric carbon emissions, biodiversity, and ecosystem productivity, accurate long-term reporting of Land-Use and Land-Cover (LULC) change in the pre-satellite era (<1972) is an imperative. Here, we used a combination of historical (1958) aerial photography and contemporary remote sensing data to map long-term changes in the extent and structure of the tropical forest surrounding Yangambi (DR Congo) in the central Congo Basin. Our study leveraged structure-from-motion and a convolutional neural network-based LULC classifier, using synthetic landscape-based image augmentation to map historical forest cover across a large orthomosaic (~93,431 ha) geo-referenced to ~4.7 \uffc2\uffb1 4.3 m at submeter resolution. A comparison with contemporary LULC data showed a shift from previously highly regular industrial deforestation of large areas to discrete smallholder farming clearing, increasing landscape fragmentation and providing opportunties for substantial forest regrowth. We estimated aboveground carbon gains through reforestation to range from 811 to 1592 Gg C, partially offsetting historical deforestation (2416 Gg C), in our study area. Efforts to quantify long-term canopy texture changes and their link to aboveground carbon had limited to no success. Our analysis provides methods and insights into key spatial and temporal patterns of deforestation and reforestation at a multi-decadal scale, providing a historical context for past and ongoing forest research in the area.

", "keywords": ["Agriculture and Food Sciences", "0301 basic medicine", "aerial survey", "550", "Science", "CONSERVATION", "ANTHROPOGENIC DISTURBANCE", "03 medical and health sciences", "TROPICAL DEFORESTATION", "RATES", "congo basin", "[SDE.ES]Environmental Sciences/Environment and Society", "cnn", "580", "CARBON EMISSIONS", "aerial survey; data recovery; CNN; deep learning; SfM; Congo Basin", "0303 health sciences", "PHOTOGRAPHS", "Q", "deep learning", "15. Life on land", "Congo Basin", "LAND-COVER", "sfm", "13. Climate action", "Earth and Environmental Sciences", "SfM", "cavelab", "General Earth and Planetary Sciences", "CO2", "VEGETATION", "SELF-SIMILARITY", "CNN", "data recovery"]}, "links": [{"href": "http://www.mdpi.com/2072-4292/12/4/638/pdf"}, {"href": "https://www.mdpi.com/2072-4292/12/4/638/pdf"}, {"href": "https://doi.org/10.3390/rs12040638"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Remote%20Sensing", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/rs12040638", "name": "item", "description": "10.3390/rs12040638", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/rs12040638"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2020-02-14T00:00:00Z"}}, {"id": "10.3390/rs13214486", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:21:08Z", "type": "Journal Article", "created": "2021-11-09", "title": "Deep Neural Networks to Detect Weeds from Crops in Agricultural Environments in Real-Time: A Review", "description": "

Automation, including machine learning technologies, are becoming increasingly crucial in agriculture to increase productivity. Machine vision is one of the most popular parts of machine learning and has been widely used where advanced automation and control have been required. The trend has shifted from classical image processing and machine learning techniques to modern artificial intelligence (AI) and deep learning (DL) methods. Based on large training datasets and pre-trained models, DL-based methods have proven to be more accurate than previous traditional techniques. Machine vision has wide applications in agriculture, including the detection of weeds and pests in crops. Variation in lighting conditions, failures to transfer learning, and object occlusion constitute key challenges in this domain. Recently, DL has gained much attention due to its advantages in object detection, classification, and feature extraction. DL algorithms can automatically extract information from large amounts of data used to model complex problems and is, therefore, suitable for detecting and classifying weeds and crops. We present a systematic review of AI-based systems to detect weeds, emphasizing recent trends in DL. Various DL methods are discussed to clarify their overall potential, usefulness, and performance. This study indicates that several limitations obstruct the widespread adoption of AI/DL in commercial applications. Recommendations for overcoming these challenges are summarized.

", "keywords": ["0106 biological sciences", "2. Zero hunger", "precision agriculture", "deep learning in agriculture; precision agriculture; weed detection; robotic weed control; machine vision for weed control", "Precision agriculture", "Machine vision for weed control", "robotic weed control", "weed detection", "Science", "Q", "04 agricultural and veterinary sciences", "01 natural sciences", "deep learning in agriculture", "Deep learning in agriculture", "0401 agriculture", " forestry", " and fisheries", "machine vision for weed control", "Weed detection", "Robotic weed control"]}, "links": [{"href": "https://www.mdpi.com/2072-4292/13/21/4486/pdf"}, {"href": "https://doi.org/10.3390/rs13214486"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Remote%20Sensing", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/rs13214486", "name": "item", "description": "10.3390/rs13214486", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/rs13214486"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2021-01-01T00:00:00Z"}}, {"id": "10.3390/rs14092253", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:21:09Z", "type": "Journal Article", "created": "2022-05-09", "title": "Extraction of Agricultural Fields via DASFNet with Dual Attention Mechanism and Multi-scale Feature Fusion in South Xinjiang, China", "description": "

Agricultural fields are essential in providing human beings with paramount food and other materials. Quick and accurate identification of agricultural fields from the remote sensing images is a crucial task in digital and precision agriculture. Deep learning methods have the advantages of fast and accurate image segmentation, especially for extracting the agricultural fields from remote sensing images. This paper proposed a deep neural network with a dual attention mechanism and a multi-scale feature fusion (Dual Attention and Scale Fusion Network, DASFNet) to extract the cropland from a GaoFen-2 (GF-2) image of 2017 in Alar, south Xinjiang, China. First, we constructed an agricultural field segmentation dataset from the GF-2 image. Next, seven evaluation indices were selected to assess the extraction accuracy, including the location shift, to reveal the spatial relationship and facilitate a better evaluation. Finally, we proposed DASFNet incorporating three ameliorated and novel deep learning modules with the dual attention mechanism and multi-scale feature fusion methods. The comparison of these modules indicated their effects and advantages. Compared with different segmentation convolutional neural networks, DASFNet achieved the best testing accuracy in extracting fields with an F1-score of 0.9017, an intersection over a union of 0.8932, a Kappa coefficient of 0.8869, and a location shift of 1.1752 pixels. Agricultural fields can be extracted automatedly and accurately using DASFNet, which reduces the manual record of the agricultural field information and is conducive to further farmland surveys, protection, and management.

", "keywords": ["2. Zero hunger", "agricultural field extraction", "GaoFen-2 (GF-2)", "agricultural field extraction; attention mechanism; deep learning; GaoFen-2 (GF-2); multi-scale feature fusion", "Science", "Q", "deep learning", "0401 agriculture", " forestry", " and fisheries", "04 agricultural and veterinary sciences", "attention mechanism", "multi-scale feature fusion"]}, "links": [{"href": "http://www.mdpi.com/2072-4292/14/9/2253/pdf"}, {"href": "https://doi.org/10.3390/rs14092253"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Remote%20Sensing", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/rs14092253", "name": "item", "description": "10.3390/rs14092253", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/rs14092253"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2022-05-07T00:00:00Z"}}, {"id": "10.3390/s22051851", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:21:10Z", "type": "Journal Article", "created": "2022-02-28", "title": "Embedded Temporal Convolutional Networks for Essential Climate Variables Forecasting", "description": "

Forecasting the values of essential climate variables like land surface temperature and soil moisture can play a paramount role in understanding and predicting the impact of climate change. This work concerns the development of a deep learning model for analyzing and predicting spatial time series, considering both satellite derived and model-based data assimilation processes. To that end, we propose the Embedded Temporal Convolutional Network (E-TCN) architecture, which integrates three different networks, namely an encoder network, a temporal convolutional network, and a decoder network. The model accepts as input satellite or assimilation model derived values, such as land surface temperature and soil moisture, with monthly periodicity, going back more than fifteen years. We use our model and compare its results with the state-of-the-art model for spatiotemporal data, the ConvLSTM model. To quantify performance, we explore different cases of spatial resolution, spatial region extension, number of training examples and prediction windows, among others. The proposed approach achieves better performance in terms of prediction accuracy, while using a smaller number of parameters compared to the ConvLSTM model. Although we focus on two specific environmental variables, the method can be readily applied to other variables of interest.

", "keywords": ["deep learning; time-series forecasting; remote sensing; climate variables; surface temperature; soil moisture", "Chemical technology", "Temperature", "0211 other engineering and technologies", "deep learning", "climate variables", "TP1-1185", "02 engineering and technology", "surface temperature", "time-series forecasting", "Article", "remote sensing", "Soil", "13. Climate action", "0202 electrical engineering", " electronic engineering", " information engineering", "soil moisture"]}, "links": [{"href": "http://www.mdpi.com/1424-8220/22/5/1851/pdf"}, {"href": "https://www.mdpi.com/1424-8220/22/5/1851/pdf"}, {"href": "https://doi.org/10.3390/s22051851"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Sensors", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.3390/s22051851", "name": "item", "description": "10.3390/s22051851", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.3390/s22051851"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2022-02-26T00:00:00Z"}}, {"id": "10.5220/0011265600003271", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:21:55Z", "type": "Journal Article", "created": "2022-07-16", "title": "Toward Autonomous Mobile Robot Navigation in Early-Stage Crop Growth", "description": "Open AccessPeer reviewed", "keywords": ["2. Zero hunger", "Autonomous Navigation", "Row following", "Time-of-Flight Camera", "0401 agriculture", " forestry", " and fisheries", "Deep learning", "04 agricultural and veterinary sciences", "Early-stage Crop-growth"], "contacts": [{"organization": "Emmi, Luis, Herrera-Diaz, Jesus, Gonzalez-de-Santos, Pablo,", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5220/0011265600003271"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Proceedings%20of%20the%2019th%20International%20Conference%20on%20Informatics%20in%20Control%2C%20Automation%20and%20Robotics", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.5220/0011265600003271", "name": "item", "description": "10.5220/0011265600003271", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5220/0011265600003271"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2022-01-01T00:00:00Z"}}, {"id": "10.5281/zenodo.11200024", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:22:14Z", "type": "Dataset", "title": "Segmentation of particulate organic matter in X-ray Computed Tomography images of soil aggregates with deep convolutional networks", "description": "unspecifiedThis is a dataset that accompanies the paper entitled \u2018Segmentation of particulate organic matter in X-ray Computed Tomography images of soil aggregates with deep convolutional networks\u2019 by Oliveira, A.B., Bordonal, R.O., Peixinho, A.Z., Carvalho, J.L.N., Ferreira, T.R. The files will be publicly accessible when the paper is published.", "keywords": ["Soil carbon stability", "Image processing", "Interactive machine learning", "Deep learning", "Synchrotron"], "contacts": [{"organization": "Oliveira, Aline, Bordonal, Ricardo, Peixinho, Alan, Carvalho, Jo\u00e3o Luis, Ferreira, Talita,", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5281/zenodo.11200024"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.11200024", "name": "item", "description": "10.5281/zenodo.11200024", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.11200024"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2024-01-01T00:00:00Z"}}, {"id": "10.5281/zenodo.15017580", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:22:55Z", "type": "Report", "title": "Analysis of tomato sepal fungal infection using hyperspectral imaging and deep learning", "description": "Abstract of the research on detection of fungal infection on tomato sepals using convolutional neural networks, presented at the WHISPERS conference in Helsinki, Finland, 2024.", "keywords": ["Hyperspectral", "Infection detection", "Deep learning", "Classification", "Imaging"], "contacts": [{"organization": "Filipovi\u0107, Vladan, Grbovi\u0107, \u017deljana, Chauhan, Aneesh, de Villiers, Hendrik, Panic, Marko, Brdar, Sanja,", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5281/zenodo.15017580"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.15017580", "name": "item", "description": "10.5281/zenodo.15017580", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.15017580"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2024-12-09T00:00:00Z"}}, {"id": "10.5281/zenodo.4896828", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:23:29Z", "type": "Report", "title": "Semantic Segmentation of Tomato Sepals on Hyperspectral Images Using Deep Learning", "description": "Open Access{'references': ['Z. Grbovic, M. Panic, O. Marko, S. Brdar and V. Crnojevic: ``Wheat Ear Detection in RGB and Thermal Images Using Deep Neural Networks'', Conference on Machine Learning and Data Mining, MLDM 2019, New York, NY, USA, July 20-25, 2019, Proceedings, Volume II, pp. 875--889, ibai publishing, 2019.', 'Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. 'U-net: Convolutional networks for biomedical image segmentation. ' In International Conference on Medical image computing and computer-assisted intervention, pp. 234-241. Springer, Cham, 2015.', 'Badrinarayanan, V., Kendall, A. and Cipolla, R., 2017. Segnet: A deep convolutional encoder-decoder architecture for image segmentation. IEEE transactions on pattern analysis and machine intelligence, 39(12), pp.2481-2495']}", "keywords": ["deep learning", " hyperspectral imaging", " semantic segmentation", " tomato sepals"], "contacts": [{"organization": "\u017deljana Grbovi\u0107, Brki\u0107, Milica, Pani\u0107, Marko, Brdar, Sanja, Echtelt, Esther Hogeveen-Van, Chauhan, Aneesh,", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5281/zenodo.4896828"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.4896828", "name": "item", "description": "10.5281/zenodo.4896828", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.4896828"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2021-01-01T00:00:00Z"}}, {"id": "10.5281/zenodo.4896827", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:23:29Z", "type": "Report", "title": "Semantic Segmentation of Tomato Sepals on Hyperspectral Images Using Deep Learning", "description": "Open Access{'references': ['Z. Grbovic, M. Panic, O. Marko, S. Brdar and V. Crnojevic: ``Wheat Ear Detection in RGB and Thermal Images Using Deep Neural Networks'', Conference on Machine Learning and Data Mining, MLDM 2019, New York, NY, USA, July 20-25, 2019, Proceedings, Volume II, pp. 875--889, ibai publishing, 2019.', 'Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. 'U-net: Convolutional networks for biomedical image segmentation. ' In International Conference on Medical image computing and computer-assisted intervention, pp. 234-241. Springer, Cham, 2015.', 'Badrinarayanan, V., Kendall, A. and Cipolla, R., 2017. Segnet: A deep convolutional encoder-decoder architecture for image segmentation. IEEE transactions on pattern analysis and machine intelligence, 39(12), pp.2481-2495']}", "keywords": ["deep learning", " hyperspectral imaging", " semantic segmentation", " tomato sepals"], "contacts": [{"organization": "\u017deljana Grbovi\u0107, Brki\u0107, Milica, Pani\u0107, Marko, Brdar, Sanja, Echtelt, Esther Hogeveen-Van, Chauhan, Aneesh,", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5281/zenodo.4896827"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.4896827", "name": "item", "description": "10.5281/zenodo.4896827", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.4896827"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2021-01-01T00:00:00Z"}}, {"id": "10.5281/zenodo.5615357", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:23:31Z", "type": "Dataset", "title": "Supplementary Table for Earth observation data-driven cropland soil monitoring: A review", "description": "Table including 46 manuscripts written in English referring to topsoil monitoring related to Earth observation data-driven cropland soil monitoring: A review paper.", "keywords": ["soil organic carbon", "hyperspectral", "spectral signatures", "carbon farming", "deep learning", "earth observation", "food security", "15. Life on land", "common agricultural policy"], "contacts": [{"organization": "Tziolas, Nikolaos, Tsakiridis, Nikolaos, Chabrillat, Sabine, Dematt\u00ea, Jos\u00e9 A.M., Ben-Dor, Eyal, Gholizadeh, Asa, Zalidis, George, Van Wesemael, Bas,", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5281/zenodo.5615357"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.5615357", "name": "item", "description": "10.5281/zenodo.5615357", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.5615357"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2021-10-22T00:00:00Z"}}, {"id": "10.5281/zenodo.7257773", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:23:41Z", "type": "Report", "title": "DEEP LEARNING INSTANCE LEVEL SEGMENTATION OF TOMATO SEPALS ON HYPERSPECTRAL IMAGES", "description": "Tomatoes are extremely prone to pathogenic fungi infections. During product storage, inadequate temperature and humidity conditions can create ideal settings for the fungi to germinate. The most fragile parts of the tomatoes are the sepals, especially on their tips, which are the main entrance spots for fungal spores. Segmentation of tomato sepals in hyperspectral images is an essential step in designing automated systems for fungal infection sensitivity assessment. Precise segmentation of sepals could contribute to predictive models for early assessment of the risk of undesired fungal occurrence.
In this study, we analyze the state-of-the-art deep learning architectures for instance segmentation to identify which architecture provides the most precise tomato sepal segmentation in hyperspectral images.", "keywords": ["hyperspectral images", "instance segmentation", "deep learning"], "contacts": [{"organization": "\u017deljana Grbovi\u0107, Marko Pani\u0107, Sanja Brdar, Esther Hogeveen-van Echtelt, Manon Mensink, Ernst Woltering Woltering, Aneesh Chauhan,", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5281/zenodo.7257773"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.7257773", "name": "item", "description": "10.5281/zenodo.7257773", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.7257773"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2022-09-26T00:00:00Z"}}, {"id": "10.5281/zenodo.7257774", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:23:41Z", "type": "Report", "title": "DEEP LEARNING INSTANCE LEVEL SEGMENTATION OF TOMATO SEPALS ON HYPERSPECTRAL IMAGES", "description": "Tomatoes are extremely prone to pathogenic fungi infections. During product storage, inadequate temperature and humidity conditions can create ideal settings for the fungi to germinate. The most fragile parts of the tomatoes are the sepals, especially on their tips, which are the main entrance spots for fungal spores. Segmentation of tomato sepals in hyperspectral images is an essential step in designing automated systems for fungal infection sensitivity assessment. Precise segmentation of sepals could contribute to predictive models for early assessment of the risk of undesired fungal occurrence.
In this study, we analyze the state-of-the-art deep learning architectures for instance segmentation to identify which architecture provides the most precise tomato sepal segmentation in hyperspectral images.", "keywords": ["hyperspectral images", "instance segmentation", "deep learning"], "contacts": [{"organization": "Grbovi\u0107, \u017deljana, Pani\u0107, Marko, Brdar, Sanja, Echtelt, Esther Hogeveen-Van, Mensink, Manon, Woltering, Ernst Woltering, Chauhan, Aneesh,", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5281/zenodo.7257774"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.7257774", "name": "item", "description": "10.5281/zenodo.7257774", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.7257774"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2022-09-26T00:00:00Z"}}, {"id": "10.5281/zenodo.8089856", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:23:47Z", "type": "Journal Article", "created": "2020-08-03", "title": "Source localization in resource-constrained sensor networks based on deep learning", "description": "Source localization with a network of low-cost motes with limited processing, memory, and energy resources is considered in this paper. The state-of-the-art methods are mostly based on complicated signal processing approaches in which motes send their (processed) data to a fusion center (FC) wherein the source is localized. These methods are resource-demanding and mostly do not meet the limitations of motes and network. In this paper, we consider distributed detection where each mote performs a binary hypothesis test to detect locally the existence of a desired source and sends its (potentially erroneous) decision to FC during just one bit (1 indicates source existence and 0 otherwise). Hence, both processing and bandwidth constraints are met. We propose to use an artificial neural network (ANN) to correct erroneous local decisions. After error correction, the region affected by the source is specified by nodes with decision 1. Moreover, we propose to localize the source by deep learning in FC which converts the network of decisions 1 and 0 to a black and white image with white pixels in the locations of motes with decision 1. The proposed schemes of error correction by ANN (ECANN) and source localization with deep learning (SoLDeL) were evaluated in a fire detection application. We showed that SoLDeL performs appropriately and scales well into large networks. Moreover, the applicability of ECANN in delineation of farm management zones was illustrated.", "keywords": ["Artificial neural network (ANN)", "Internet of things (IoT)", "0202 electrical engineering", " electronic engineering", " information engineering", "Deep learning", "Target tracking", "Error type II", "02 engineering and technology", "Decentralized detection", "15. Life on land", "Wireless sensor networks (WSN)", "Error type I", "Source localization"]}, "links": [{"href": "https://link.springer.com/content/pdf/10.1007/s00521-020-05253-3.pdf"}, {"href": "https://doi.org/10.5281/zenodo.8089856"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Neural%20Computing%20and%20Applications", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.8089856", "name": "item", "description": "10.5281/zenodo.8089856", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.8089856"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2020-08-03T00:00:00Z"}}, {"id": "10.5281/zenodo.8091934", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:23:48Z", "type": "Journal Article", "created": "2022-08-18", "title": "Data mining of urban soil spectral library for estimating organic carbon", "description": "Accurate quantification of urban soil organic carbon (SOC) is essential for understanding anthropogenic changes and further guiding effective city managements. Visible and near infrared (vis\u2013NIR) spectroscopy can monitor the SOC content in a time- and cost-effective manner. However, processes and mechanisms dominating the relationships between SOC and spectral data in urban soils remain unknown. The main objective of this paper was to evaluate whether multiple stratification strategies (i.e., based on land-use/land-cover [LULC], pH, and spectral clustering) resulted in better predicted performance for SOC compared to the non-stratified (global) models. Results showed that regarding the non-stratified models, the convolutional neural network (CNN) model exhibited the best performance (validation R2 = 0.73), followed by Cubist (validation R2 = 0.66) and memorybased learning (validation R2 = 0.65). After LULC stratification, Cubist model achieved the best prediction (validation R2 = 0.76), improving the value of ratio of performance to interquartile distance by 0.11 compared to the global CNN model. Areas with high SOC values were mainly located in the city center. Stratification by LULC class is a promising strategy for addressing the impact of the soil-landscape diversity and complexity on vis\u2013NIR spectral estimation of SOC in urban soil spectral library.", "keywords": ["Urban soil", "Stratified modeling", "13. Climate action", "Soil organic carbon", "11. Sustainability", "0401 agriculture", " forestry", " and fisheries", "Deep learning", "04 agricultural and veterinary sciences", "15. Life on land", "Soil spectral library"]}, "links": [{"href": "https://doi.org/10.5281/zenodo.8091934"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Geoderma", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.8091934", "name": "item", "description": "10.5281/zenodo.8091934", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.8091934"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2022-11-01T00:00:00Z"}}, {"id": "10.5281/zenodo.8092644", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:23:48Z", "type": "Journal Article", "created": "2022-05-08", "title": "Extraction of Agricultural Fields via DASFNet with Dual Attention Mechanism and Multi-scale Feature Fusion in South Xinjiang, China", "description": "

Agricultural fields are essential in providing human beings with paramount food and other materials. Quick and accurate identification of agricultural fields from the remote sensing images is a crucial task in digital and precision agriculture. Deep learning methods have the advantages of fast and accurate image segmentation, especially for extracting the agricultural fields from remote sensing images. This paper proposed a deep neural network with a dual attention mechanism and a multi-scale feature fusion (Dual Attention and Scale Fusion Network, DASFNet) to extract the cropland from a GaoFen-2 (GF-2) image of 2017 in Alar, south Xinjiang, China. First, we constructed an agricultural field segmentation dataset from the GF-2 image. Next, seven evaluation indices were selected to assess the extraction accuracy, including the location shift, to reveal the spatial relationship and facilitate a better evaluation. Finally, we proposed DASFNet incorporating three ameliorated and novel deep learning modules with the dual attention mechanism and multi-scale feature fusion methods. The comparison of these modules indicated their effects and advantages. Compared with different segmentation convolutional neural networks, DASFNet achieved the best testing accuracy in extracting fields with an F1-score of 0.9017, an intersection over a union of 0.8932, a Kappa coefficient of 0.8869, and a location shift of 1.1752 pixels. Agricultural fields can be extracted automatedly and accurately using DASFNet, which reduces the manual record of the agricultural field information and is conducive to further farmland surveys, protection, and management.

", "keywords": ["2. Zero hunger", "agricultural field extraction", "GaoFen-2 (GF-2)", "agricultural field extraction; attention mechanism; deep learning; GaoFen-2 (GF-2); multi-scale feature fusion", "Science", "Q", "deep learning", "0401 agriculture", " forestry", " and fisheries", "04 agricultural and veterinary sciences", "attention mechanism", "multi-scale feature fusion"]}, "links": [{"href": "http://www.mdpi.com/2072-4292/14/9/2253/pdf"}, {"href": "https://doi.org/10.5281/zenodo.8092644"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/Remote%20Sensing", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "10.5281/zenodo.8092644", "name": "item", "description": "10.5281/zenodo.8092644", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5281/zenodo.8092644"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2022-05-07T00:00:00Z"}}, {"id": "3132124173", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:26:19Z", "type": "Journal Article", "created": "2021-02-22", "title": "Infrared Metasurface Augmented by Deep Learning for Monitoring Dynamics between All Major Classes of Biomolecules", "description": "Abstract

Insights into the fascinating molecular world of biological processes are crucial for understanding diseases, developing diagnostics, and effective therapeutics. These processes are complex as they involve interactions between four major classes of biomolecules, i.e., proteins, nucleic acids, carbohydrates, and lipids, which makes it important to be able to discriminate between all these different biomolecular species. In this work, a deep learning\uffe2\uff80\uff90augmented, chemically\uffe2\uff80\uff90specific nanoplasmonic technique that enables such a feat in a label\uffe2\uff80\uff90free manner to not disrupt native processes is presented. The method uses a highly sensitive multiresonant plasmonic metasurface in a microfluidic device, which enhances infrared absorption across a broadband mid\uffe2\uff80\uff90IR spectrum and in water, despite its strongly overlapping absorption bands. The real\uffe2\uff80\uff90time format of the optofluidic method enables the collection of a vast amount of spectrotemporal data, which allows the construction of a deep neural network to discriminate accurately between all major classes of biomolecules. The capabilities of the new method are demonstrated by monitoring of a multistep bioassay containing sucrose\uffe2\uff80\uff90 and nucleotides\uffe2\uff80\uff90loaded liposomes interacting with a small, lipid membrane\uffe2\uff80\uff90perforating peptide. It is envisioned that the presented technology will impact the fields of biology, bioanalytics, and pharmacology from fundamental research and disease diagnostics to drug development.Soils store more carbon than other terrestrial ecosystems1,2. How soil organic carbon (SOC) forms and persists remains uncertain1,3, which makes it challenging to understand how it will respond to climatic change3,4. It has been suggested that soil microorganisms play an important role in SOC formation, preservation and loss5\uffe2\uff80\uff937. Although microorganisms affect the accumulation and loss of soil organic matter through many pathways4,6,8\uffe2\uff80\uff9311, microbial carbon use efficiency (CUE) is an integrative metric that can capture the balance of these processes12,13. Although CUE has the potential to act as a predictor of variation in SOC storage, the role of CUE in SOC persistence remains unresolved7,14,15. Here we examine the relationship between CUE and the preservation of SOC, and interactions with climate, vegetation and edaphic properties, using a combination of global-scale datasets, a microbial-process explicit model, data assimilation, deep learning and meta-analysis. We find that CUE is at least four times as important as other evaluated factors, such as carbon input, decomposition or vertical transport, in determining SOC storage and its spatial variation across the globe. In addition, CUE shows a positive correlation with SOC content. Our findings point to microbial CUE as a major determinant of global SOC storage. Understanding the microbial processes underlying CUE and their environmental dependence may help the prediction of SOC feedback to a changing climate.Tomato is an important commercial product which is perishable by nature and highly susceptible to fungal incidence once it is harvested. Not all tomatoes are equally vulnerable to pathogenic fungi, and an early detection of the vulnerable ones can help in taking timely preventive actions, ranging from isolating tomato batches to adjusting storage conditions, but also in making right business decisions like dynamic pricing based on quality or better shelf life estimate. More importantly, early detection of vulnerable produce can help in taking timely actions to minimize potential post-harvest losses. This paper investigates Near-infrared (NIR) hyperspectral imaging (1000\uffe2\uff80\uff931700\uffc2\uffa0nm) and machine learning to build models to automatically predict the susceptibility of sepals of recently harvested tomatoes to future fungal infections. Hyperspectral images of newly harvested tomatoes (cultivar Brioso) from 5 different growers were acquired before the onset of any visible fungal infection. After imaging, the tomatoes were placed under controlled conditions suited for fungal germination and growth for a 4-day period, and then imaged using normal color cameras. All sepals in the color images were ranked for fungal severity using crowdsourcing, and the final severity of each sepal was fused using principal component analysis. A novel hyperspectral data processing pipeline is presented which was used to automatically segment the tomato sepals from spectral images with multiple tomatoes connected via a truss. The key modelling question addressed in this research is whether there is a correlation between the hyperspectral data captured at harvest and the fungal infection observed 4 days later. Using 10-fold and group k-fold cross-validation, XG-Boost and Random Forest based regression models were trained on the features derived from the hyperspectral data corresponding to each sepal in the training set and tested on hold out test set. The best model found a Pearson correlation of 0.837, showing that there is strong linear correlation between the NIR spectra and the future fungal severity of the sepal. The sepal specific predictions were aggregated to predict the susceptibility of individual tomatoes, and a correlation of 0.92 was found. Besides modelling, focus is also on model interpretation, particularly to understand which spectral features are most relevant to model prediction. Two approaches to model interpretation were explored, feature importance and SHAP (SHapley Additive exPlanations), resulting in similar conclusions that the NIR range between 1390\uffe2\uff80\uff931420\uffc2\uffa0nm contributes most to the model\uffe2\uff80\uff99s final decision.

The discovery and development of new drugs are extremely long and costly processes. Recent progress in artificial intelligence has made a positive impact on the drug development pipeline. Numerous challenges have been addressed with the growing exploitation of drug-related data and the advancement of deep learning technology. Several model frameworks have been proposed to enhance the performance of deep learning algorithms in molecular design. However, only a few have had an immediate impact on drug development since computational results may not be confirmed experimentally. This systematic review aims to summarize the different deep learning architectures used in the drug discovery process and are validated with further in vivo experiments. For each presented study, the proposed molecule or peptide that has been generated or identified by the deep learning model has been biologically evaluated in animal models. These state-of-the-art studies highlight that even if artificial intelligence in drug discovery is still in its infancy, it has great potential to accelerate the drug discovery cycle, reduce the required costs, and contribute to the integration of the 3R (Replacement, Reduction, Refinement) principles. Out of all the reviewed scientific articles, seven algorithms were identified: recurrent neural networks, specifically, long short-term memory (LSTM-RNNs), Autoencoders (AEs) and their Wasserstein Autoencoders (WAEs) and Variational Autoencoders (VAEs) variants; Convolutional Neural Networks (CNNs); Direct Message Passing Neural Networks (D-MPNNs); and Multitask Deep Neural Networks (MTDNNs). LSTM-RNNs were the most used architectures with molecules or peptide sequences as inputs.

", "keywords": ["Deep Learning", "Artificial Intelligence", "Drug Discovery", "Review", "Neural Networks", " Computer", "Algorithms", "3. Good health"]}, "links": [{"href": "https://www.mdpi.com/1422-0067/24/7/6573/pdf"}, {"href": "https://doi.org/PMC10095548"}, {"rel": "related", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/International%20Journal%20of%20Molecular%20Sciences", "name": "related record", "description": "related record", "type": "application/json"}, {"rel": "self", "type": "application/geo+json", "title": "PMC10095548", "name": "item", "description": "PMC10095548", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/PMC10095548"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2023-03-31T00:00:00Z"}}, {"id": "PMC11469153", "type": "Feature", "geometry": null, "properties": {"updated": "2026-07-25T16:28:07Z", "type": "Journal Article", "created": "2021-02-22", "title": "Infrared Metasurface Augmented by Deep Learning for Monitoring Dynamics between All Major Classes of Biomolecules", "description": "Abstract

Insights into the fascinating molecular world of biological processes are crucial for understanding diseases, developing diagnostics, and effective therapeutics. These processes are complex as they involve interactions between four major classes of biomolecules, i.e., proteins, nucleic acids, carbohydrates, and lipids, which makes it important to be able to discriminate between all these different biomolecular species. In this work, a deep learning\uffe2\uff80\uff90augmented, chemically\uffe2\uff80\uff90specific nanoplasmonic technique that enables such a feat in a label\uffe2\uff80\uff90free manner to not disrupt native processes is presented. The method uses a highly sensitive multiresonant plasmonic metasurface in a microfluidic device, which enhances infrared absorption across a broadband mid\uffe2\uff80\uff90IR spectrum and in water, despite its strongly overlapping absorption bands. The real\uffe2\uff80\uff90time format of the optofluidic method enables the collection of a vast amount of spectrotemporal data, which allows the construction of a deep neural network to discriminate accurately between all major classes of biomolecules. The capabilities of the new method are demonstrated by monitoring of a multistep bioassay containing sucrose\uffe2\uff80\uff90 and nucleotides\uffe2\uff80\uff90loaded liposomes interacting with a small, lipid membrane\uffe2\uff80\uff90perforating peptide. It is envisioned that the presented technology will impact the fields of biology, bioanalytics, and pharmacology from fundamental research and disease diagnostics to drug development.Tomato is an important commercial product which is perishable by nature and highly susceptible to fungal incidence once it is harvested. Not all tomatoes are equally vulnerable to pathogenic fungi, and an early detection of the vulnerable ones can help in taking timely preventive actions, ranging from isolating tomato batches to adjusting storage conditions, but also in making right business decisions like dynamic pricing based on quality or better shelf life estimate. More importantly, early detection of vulnerable produce can help in taking timely actions to minimize potential post-harvest losses. This paper investigates Near-infrared (NIR) hyperspectral imaging (1000\uffe2\uff80\uff931700\uffc2\uffa0nm) and machine learning to build models to automatically predict the susceptibility of sepals of recently harvested tomatoes to future fungal infections. Hyperspectral images of newly harvested tomatoes (cultivar Brioso) from 5 different growers were acquired before the onset of any visible fungal infection. After imaging, the tomatoes were placed under controlled conditions suited for fungal germination and growth for a 4-day period, and then imaged using normal color cameras. All sepals in the color images were ranked for fungal severity using crowdsourcing, and the final severity of each sepal was fused using principal component analysis. A novel hyperspectral data processing pipeline is presented which was used to automatically segment the tomato sepals from spectral images with multiple tomatoes connected via a truss. The key modelling question addressed in this research is whether there is a correlation between the hyperspectral data captured at harvest and the fungal infection observed 4 days later. Using 10-fold and group k-fold cross-validation, XG-Boost and Random Forest based regression models were trained on the features derived from the hyperspectral data corresponding to each sepal in the training set and tested on hold out test set. The best model found a Pearson correlation of 0.837, showing that there is strong linear correlation between the NIR spectra and the future fungal severity of the sepal. The sepal specific predictions were aggregated to predict the susceptibility of individual tomatoes, and a correlation of 0.92 was found. Besides modelling, focus is also on model interpretation, particularly to understand which spectral features are most relevant to model prediction. Two approaches to model interpretation were explored, feature importance and SHAP (SHapley Additive exPlanations), resulting in similar conclusions that the NIR range between 1390\uffe2\uff80\uff931420\uffc2\uffa0nm contributes most to the model\uffe2\uff80\uff99s final decision.