Search

Early detection and diagnosis of crop anomalies is crucial for enhancing crop yield and quality. Recently, the combination of machine learning and deep learning with hyperspectral images has significantly improved the efficiency of crop detection. However, acquiring a large amount of properly annotated hyperspectral data on stressed crops requires extensive biochemical experiments and specialized knowledge. This limitation poses a challenge to the construction of large-scale datasets for crop stress analysis. Meta-learning is a learning approach that is capable of learning to learn and can achieve high detection accuracy with limited training samples. In this paper, we introduce meta-learning to hyperspectral imaging and crop detection for the first time. In addition, we gathered 88 hyperspectral images of drought-stressed tomato plants and 68 images of freeze-stressed tomato plants. The data related to drought serve as the source domain, while the data related to frost damage serve as the target domain. Due to the difficulty of obtaining target domain data from real-world testing scenarios, only a limited amount of target domain data and source domain data are used for model training. The results indicated that meta-learning, with a minimum of eight target domain samples, achieved a detection accuracy of 69.57%, precision of 59.29%, recall of 66.32% and F 1 -score of 62.61% for classifying the severity of frost stress, surpassing other methods with a target domain sample size of 20. Moreover, for determining whether the plants were under stress, meta-learning, with a minimum of four target domain samples, achieved a detection accuracy of 89.1%, precision of 89.72%, recall of 93.08% and F 1 -score of 91.37% outperforming other methods at a target domain sample size of 20. The results show that meta-learning methods require significantly less data across different domains compared to other methods. The performance of meta-learning techniques thoroughly demonstrates the feasibility of rapidly detecting crop stress without the need for collecting a large amount of target stress data. This research alleviates the data annotation pressure for researchers and provides a foundation for detection personnel to anticipate and prevent potential large-scale stress damage to crops. [ABSTRACT FROM AUTHOR]

Cotton is one of the most important cash crops in Xinjiang, and timely seedling inspection and replenishment at the seedling stage are essential for cotton's late production management and yield formation. The background conditions of the cotton seedling stage are complex and variable, and deep learning methods are widely used to extract target objects from the complex background. Therefore, this study takes seedling cotton as the research object and uses three deep learning algorithms, YOLOv5, YOLOv7, and CenterNet, for cotton seedling detection and counting using images at six different times of the cotton seedling period based on multispectral images collected by UAVs to develop a model applicable to the whole cotton seedling period. The results showed that when tested with data collected at different times, YOLOv7 performed better overall in detection and counting, and the T4 dataset performed better in each test set. Precision, Recall, and F1-Score values with the best test results were 96.9%, 96.6%, and 96.7%, respectively, and the R2, RMSE, and RRMSE indexes were 0.94, 3.83, and 2.72%, respectively. In conclusion, the UAV multispectral images acquired about 23 days after cotton sowing (T4) with the YOLOv7 algorithm achieved rapid and accurate seedling detection and counting throughout the cotton seedling stage. [ABSTRACT FROM AUTHOR]

• Proposing a framework for multi-plant phenotype extraction with minimal point clouds. • Non-parametric strategy enhances point cloud spatial representation capability. • Creating a labeled cotton point cloud dataset with multi-temporal and multi-species. Rapid and precise 3D organ segmentation is crucial for the automatic extraction of phenotypic traits, forming a fundamental prerequisite for intelligent plant breeding. The advancement of deep learning technology has replaced labor-intensive manual measurements and traditional computer vision methods, which are sensitive to parameters, in phenotypic trait extraction. However, current larger network structures not only require extensive point cloud data but also consume substantial computational resources, rendering them unsuitable for agricultural tasks with limited plant samples. Therefore, this study developed a lightweight 3D deep learning network (PEPNet) that achieves precise plant organ segmentation and stem-leaf phenotypic trait extraction. The adopted simple-but-effective network architecture and innovative modern operations, including a high-dimensional feature mapping strategy for preprocessing input points, a local feature extraction module based on inverted residual bottleneck block, and a cost-free attention block for spatial feature fusion, effectively implement multi-scale hierarchies and adaptively reduce computational overheads. Experimental results from cotton stem-leaf segmentation demonstrated that PEPNet not only presented approximately 2 × faster inference speed (9.59 ms) and throughput (146.32 plants per second) but also achieved competitive segmentation performance compared to other six state-of-the-art deep learning networks, namely PointNet++, DGCNN, CurveNet, Point Cloud Transformer, PointMLP, and SPoTr, achieving 95.99 %, 94.66 %, 95.32 %, and 91.31 % in Precision, Recall, F1-score, and mIoU, respectively. In transferability experiments with tomato and soybean plants, PEPNet achieved almost all the best metrics and significantly outperformed the second-best model (CurveNet). Furthermore, ablation study verified the optimal trade-off between efficiency and accuracy in this network. Any modifications to the modules could potentially disrupt the optimal trade-off. This work could contribute to reducing computational resources and annotation costs for applying segmentation methods in high-throughput phenotyping tasks. [ABSTRACT FROM AUTHOR]

Background: The aim of this study was to investigate the influence of convolution kernel and iterative reconstruction on the diagnostic performance of radiomics and deep learning (DL) in lung adenocarcinomas. Methods: A total of 183 patients with 215 lung adenocarcinomas were included in this study. All CT imaging data was reconstructed with three reconstruction algorithms (ASiR at 0%, 30%, 60% strength), each with two convolution kernels (bone and standard). A total of 171 nodules were selected as the training‐validation set, whereas 44 nodules were selected as the testing set. Logistic regression and a DL framework‐DenseNets were selected to tackle the task. Three logical experiments were implemented to fully explore the influence of the studied parameters on the diagnostic performance. The receiver operating characteristic curve (ROC) was used to evaluate the performance of constructed models. Results: In Experiments A and B, no statistically significant results were found in the radiomic method, whereas two and six pairs were statistically significant (P < 0.05) in the DL method. In Experiment_C, significant differences in one and four models were found in the radiomics and DL methods, respectively. Moreover, models constructed with standard convolution kernel data outperformed that constructed with bone convolution kernel data in all studied ASiR levels in the DL method. In the DL method, B0 and S60 performed best in bone and standard convolution kernel, respectively. Conclusion: The results demonstrated that DL was more susceptible to CT parameter variability than radiomics. Standard convolution kernel images seem to be more appropriate for imaging analysis. Further investigation with a larger sample size is needed. [ABSTRACT FROM AUTHOR]

To develop a deep learning system based on 3D convolutional neural networks (CNNs), and to automatically predict EGFR‐mutant pulmonary adenocarcinoma in CT images. A dataset of 579 nodules with EGFR mutation status labels of mutant (Mut) or wild‐type (WT) was retrospectively analyzed. A deep learning system, namely 3D DenseNets, was developed to process 3D patches of nodules from CT data, and learn strong representations with supervised end‐to‐end training. The 3D DenseNets were trained with a training subset of 348 nodules and tuned with a development subset of 116 nodules. A strong data augmentation technique, mixup, was used for better generalization. We evaluated our model on a holdout subset of 115 nodules. An independent public dataset of 37 nodules from the cancer imaging archive (TCIA) was also used to test the generalization of our method. Conventional radiomics analysis was also performed for comparison. Our method achieved promising performance on predicting EGFR mutation status, with AUCs of 75.8% and 75.0% for our holdout test set and public test set, respectively. Moreover, strong relations were found between deep learning feature and conventional radiomics, while deep learning worked through an enhanced radiomics manner, that is, deep learned radiomics (DLR), in terms of robustness, compactness and expressiveness. The proposed deep learning system predicts EGFR‐mutant of lung adenocarcinomas in CT images noninvasively and automatically, indicating its potential to help clinical decision‐making by identifying eligible patients of pulmonary adenocarcinoma for EGFR‐targeted therapy. [ABSTRACT FROM AUTHOR]

Cotton seed purity is a critical factor influencing the cotton yield. In this study, near-infrared hyperspectral imaging was used to identify seven varieties of cotton seeds. Score images formed by pixel-wise principal component analysis (PCA) showed that there were differences among different varieties of cotton seeds. Effective wavelengths were selected according to PCA loadings. A self-design convolution neural network (CNN) and a Residual Network (ResNet) were used to establish classification models. Partial least squares discriminant analysis (PLS-DA), logistic regression (LR) and support vector machine (SVM) were used as direct classifiers based on full spectra and effective wavelengths for comparison. Furthermore, PLS-DA, LR and SVM models were used for cotton seeds classification based on deep features extracted by self-design CNN and ResNet models. LR and PLS-DA models using deep features as input performed slightly better than those using full spectra and effective wavelengths directly. Self-design CNN based models performed slightly better than ResNet based models. Classification models using full spectra performed better than those using effective wavelengths, with classification accuracy of calibration, validation and prediction sets all over 80% for most models. The overall results illustrated that near-infrared hyperspectral imaging with deep learning was feasible to identify cotton seed varieties. [ABSTRACT FROM AUTHOR]

Processing tomato (Lycopersicon esculentum Mill.) is rich in vitamins and lycopene, which is favored by consumers. In this study, near-infrared hyperspectral imaging (HSI) technology (980–1660 nm) was used to detect the firmness, soluble solids, lycopene, and titratable acid content of processing tomatoes and to classify fruits at three maturity stages. Savitzky-Golay (SG) smoothing was used to reduce the noise of hyperspectral images. The average spectrum of the tomato fruit was extracted for model development. Random forest (RF), partial least squares (PLS), and recurrent neural network (RNN) were used to develop models for predicting the four quality attributes and identifying the maturity level. Results showed that the RNN model had a classification accuracy of 40% higher than RF and 17% higher than PLS. In the prediction of quality parameters, RNN models had the highest R2 value (>0.87), followed by PLS and RF models. Important wavelengths were identified by calculating its contribution values and were used to interpret the model. The results illustrated that near-infrared hyperspectral imaging technology combined with deep learning could effectively predict the quality and maturity of processing tomatoes. The work can provide a perspective on the application of HSI as a nondestructive testing approach for other agricultural products. • Use of hyperspectral imaging to predict multi-quality of processing tomatoes. • Use of hyperspectral imaging to classify maturity of processing tomatoes. • Comparison of recurrent neural networks with traditional machine learning methods. • Identification of important wavelengths in contributing to model performance. [ABSTRACT FROM AUTHOR]

• Coarse-to-fine segmentation. The proposed interaction of supervised and unsupervised models can realize fine-grained foreground segmentation. • Unsupervised background model updates. The unsupervised statistical background model can update and avoid deadlock by using segmented masks as external selectiveupdating cues. • This method is more flexible than deep learning-based methods that depend on scenespecific training. Compared with unsupervised models, it reduces the number of training samples and utilizes training datasets with no human intervention. In this paper 1 1 Dong Liang, Bin Kang and Xinyu Liu contributed equally to this work. , we investigate cross-scene video foreground segmentation via supervised and unsupervised model communication. Traditional unsupervised background subtraction methods often face the challenging problem of updating the statistical background model online. In contrast, supervised foreground segmentation methods, such as those that are based on deep learning, rely on large amounts of training data, thereby limiting their cross-scene performance. Our method leverages segmented masks from a cross-scene trained deep model (spatio-temporal attention model (STAM), pyramid scene parsing network (PSPNet), or DeepLabV3+) to seed online updates for the statistical background model (CPB), thereby refining the foreground segmentation. More flexible than methods that require scene-specific training and more data-efficient than unsupervised models, our method outperforms state-of-the-art approaches on CDNet2014, WallFlower, and LIMU according to our experimental results. The proposed framework can be integrated into a video surveillance system in a plug-and-play form to realize cross-scene foreground segmentation. [ABSTRACT FROM AUTHOR]

• Near-infrared hyperspectral imaging was used to identify coated maize varieties. • Conventional and deep learning methods were used to build classification models. • Feature extraction and fusion were evaluated and compared with good performances. • Deep learning based data analysis was efficient to identify coated maize varieties. Tracing the varieties of seeds is of great importance for the seed industry. Maize kernels for planting are generally coated to protect kernels from fungi and insects. In this study, near-infrared hyperspectral imaging ranging from 874 nm to 1734 nm was used to identify the varieties of coated maize kernels. Spectral data were extracted and preprocessed. Logistic regression (LR) and support vector machine (SVM), convolutional neural network (CNN), recurrent neural network (RNN) and Long Short-Term Memory (LSTM) were used to build classification models. Furthermore, principal component analysis (PCA), CNN, RNN and LSTM were adopted to extract features. The extracted features were fused as the inputs of the classification models. Classification models using full spectra, extracted features and fused features obtained performances with the classification accuracy over 90% in the calibration, validation and prediction sets of most models. Models using extracted features obtained equivalently or slightly worse results than those using full spectra. The models using fused features all obtained good performances, with the classification accuracy over 90% in all sets. The overall results illustrated that near-infrared hyperspectral imaging with deep learning methods was a useful alternative for identifying coated maize varieties. [ABSTRACT FROM AUTHOR]