Search Results (1,529)

Cross-scene classification of hyperspectral images poses significant challenges due to the lack of a priori knowledge and the differences in data distribution across scenes. While traditional studies have had limited use of a priori knowledge from other modalities, recent advancements in pre-trained large-scale language-vision models have shown strong performance on various downstream tasks, highlighting the potential of cross-modal assisted learning. In this paper, we propose a Semantic-aware Collaborative Parallel Network (SCPNet) to mitigate the impact of data distribution differences by incorporating linguistic modalities to assist in learning cross-domain invariant representations of hyperspectral images. SCPNet uses a parallel architecture consisting of a spatial–spectral feature extraction module and a multiscale feature extraction module, designed to capture rich image information during the feature extraction phase. The extracted features are then mapped into an optimized semantic space, where improved supervised contrastive learning clusters image features from the same category together while separating those from different categories. Semantic space bridges the gap between visual and linguistic modalities, enabling the model to mine cross-domain invariant representations from the linguistic modality. Experimental results demonstrate that SCPNet significantly outperforms existing methods on three publicly available datasets, confirming its effectiveness for cross-scene hyperspectral image classification tasks. Full article

Background/Objective: Esophageal cancer (EC) is a major global health issue due to its high mortality rate, as patients are often diagnosed at advanced stages. This research examines whether the Spectrum-Aided Vision Enhancer (SAVE), a hyperspectral imaging (HSI) technique, enhances endoscopic image categorization for superior diagnostic outcomes compared to traditional White Light Imaging (WLI) and Narrow Band Imaging (NBI). Methods: A dataset including 2400 photos categorized into eight disease types from National Taiwan University Hospital Yun-Lin Branch was utilized. Multiple machine learning and deep learning models were developed, including logistic regression, VGG16, YOLOv8, and MobileNetV2. SAVE was utilized to transform WLI photos into hyperspectral representations, and band selection was executed to enhance feature extraction and improve classification outcomes. The training and evaluation of the model incorporated precision, recall, F1-score, and accuracy metrics across WLI, NBI, and SAVE modalities. Results: The research findings indicated that SAVE surpassed both NBI and WLI by achieving superior precision, recall, and F1-scores. Logistic regression and VGG16 performed with a comparable reliability to SAVE and NBI, whereas MobileNetV2 and YOLOv8 demonstrated inconsistent yet enhanced results. Overall, SAVE exhibited exceptional categorization precision and recall, showcasing impeccable performance across many models. Conclusions: This research indicates that AI hyperspectral imaging facilitates early diagnosis of esophageal diseases, hence enhancing clinical decision-making and improving patient outcomes. The amalgamation of SAVE with machine learning and deep learning models enhances diagnostic capabilities, with SAVE and NBI surpassing WLI by offering superior tissue differentiation and diagnostic accuracy. Full article

Malignant pleural effusion, commonly referred to as MPE, is a prevalent complication associated with individuals diagnosed with neoplastic disorders. The data acquired by pleural fluid cytology is beneficial for diagnostic objectives. Consequently, the initial step in the diagnostic procedure for lung cancer is the analysis of pleural effusion fluid. This research aims to provide a cutting-edge model for analyzing PE cytology images. This model utilizes a computer-aided diagnosis (CAD) system that integrates hyperspectral imaging (HSI) technology for the classification of spectral variations. Giemsa, which is one of the most popular microscopic stains, is employed to stain the samples, after which a sensitive CCD mounted on a microscope captures the images. Subsequently, the HSI model is tasked with obtaining the image spectra. Principal Component Analysis (PCA) constitutes the concluding phase in the classification procedure of various cell types. We expect that the suggested technique will enable medical professionals to stage lung cancer more rapidly. In the future, we aspire to develop an extensive data system that utilizes deep learning techniques to facilitate the automatic classification of cells, thereby ensuring the most precise diagnosis. Furthermore, enhancing accuracy and minimizing data dimensions are important priorities to accelerate diagnostics, conserve resources, and reduce computing time. Full article

Background/Objectives: Breast cancer (BC) has emerged as a significant threat among female malignancies, resulting in approximately 670,000 fatalities. The capacity to identify BC has advanced over the past two decades because of deep learning (DL), machine learning (ML), and artificial intelligence. The early detection of BC is crucial; yet, conventional diagnostic techniques, including MRI, mammography, and biopsy, are costly, time-intensive, less sensitive, incorrect, and necessitate skilled physicians. This narrative review will examine six novel imaging approaches for BC diagnosis. Methods: Optical coherence tomography (OCT) surpasses existing approaches by providing non-invasive, high-resolution imaging. Raman Spectroscopy (RS) offers detailed chemical and structural insights into cancer tissue that traditional approaches cannot provide. Photoacoustic Imaging (PAI) provides superior optical contrast, exceptional ultrasonic resolution, and profound penetration and visualization capabilities. Hyperspectral Imaging (HSI) acquires spatial and spectral data, facilitating non-invasive tissue classification with superior accuracy compared to grayscale imaging. Contrast-Enhanced Spectral Mammography (CESM) utilizes contrast agents and dual energy to improve the visualization of blood vessels, enhance patient comfort, and surpass standard mammography in sensitivity. Multispectral Imaging (MSI) enhances tissue classification by employing many wavelength bands, resulting in high-dimensional images that surpass the ultrasound approach. The imaging techniques studied in this study are very useful for diagnosing tumors, staging them, and guiding surgery. They are not detrimental to morphological or immunohistochemical analysis, which is the gold standard for diagnosing breast cancer and determining molecular characteristics. Results: These imaging modalities provide enhanced sensitivity, specificity, and diagnostic accuracy. Notwithstanding their considerable potential, the majority of these procedures are not employed in standard clinical practices. Conclusions: Validations, standardization, and large-scale clinical trials are essential for the real-time application of these approaches. The analyzed studies demonstrated that the novel modalities displayed enhanced diagnostic efficacy, with reported sensitivities and specificities often exceeding those of traditional imaging methods. The results indicate that they may assist in early detection and surgical decision-making; however, for widespread adoption, they must be standardized, cost-reduced, and subjected to extensive clinical trials. This study offers a concise summary of each methodology, encompassing the methods and findings, while also addressing the many limits encountered in the imaging techniques and proposing solutions to mitigate these issues for future applications. Full article

The accurate and non-destructive assessment of fruit ripeness is essential for post-harvest sorting and quality management. This study evaluated a meta-inspired classification framework integrating partial least squares discriminant analysis (PLS-DA) with support vector machines (SVMs) trained on latent variables (sSVM) or on class probabilities (pSVM) derived from multiple PLS-DA components. Two kiwiberry varieties, ‘Geneva’ and ‘Weiki’, were analyzed using variety-specific and combined datasets. Performance was assessed in calibration and prediction using accuracy, F₀₅, Cohen’s kappa, precision, sensitivity, specificity, and likelihood ratios. Conventional PLS-DA provided reasonably good classification, but pSVM models, particularly those with an RBF kernel (pSVM_R), consistently outperformed other approaches and ensured higher stability across all datasets. Unlike sSVMs, which were prone to overfitting, pSVM_R models achieved the highest accuracy of 92.4–96.9%, Cohen’s kappa of 84.8–93.9%, and precision of 89.1–94.2%, clearly surpassing both score-based SVM and PLS-DA. Contrasting tendencies were observed between cultivars: ‘Geneva’ models improved during prediction, while ‘Weiki’ models declined, especially in specificity. Combined datasets provided greater stability but slightly reduced peak performance than single-variety models. These findings highlight the value of probability-enriched stacking models for non-invasive ripeness discrimination, suggesting that adaptive or hybrid strategies may further enhance generalization across diverse cultivars. Full article

Microplastics are plastic particles less than five millimeters in diameter that have led to serious environmental problems, and detecting these tiny particles is crucial to understanding their distribution and impact on the soil environment. In this paper, we propose the Self-Supervised Hierarchical Dilated Transformer Network (SHDTNet), an improved hyperspectral image classification model based on self-supervised contrastive learning, for identifying and detecting microplastics in soil. Currently, most hyperspectral image classifications rely on supervised methods, which perform well with rich training samples. However, pixel labeling in soil microplastic detection scenarios is a difficult and costly task. By employing the self-supervised contrastive learning technique, SHDTNet addresses the problem of insufficient training samples for hyperspectral images of soil microplastics and also enhances the feature extraction module in contrastive learning to improve the network model’s feature extraction capability. Experiments on self-constructed hyperspectral soil microplastic image datasets demonstrate that the proposed method accurately recognizes unique microplastics in the soil environment without errors or missed detections, outperforming several currently available soil microplastic detection methods. Full article

The development of remote methods for identifying plant light stress (LS) is an urgent task in agriculture and forestry. Evergreen conifers, which experience winter light stress (WLS) annually, are ideal subjects for studying the mechanisms of light stress and developing identification methods. Proximal hyperspectral imaging (HSI) was used to identify WLS in Platycladus orientalis. Using the random forest (RF), the spectral characteristics of P. orientalis shoots were analysed and the conditions ‘Winter Light Stress’ and ‘Optimal Condition’ were classified with high accuracy. The out-of-bag (OOB) estimate of the error rate was only 0.35%. Classification of the conditions ‘Cold Stress’ and ‘Optimal Condition’—with an OOB estimate of error rate of 3.19%—can also be considered successful. The conditions ‘Winter Light Stress’ and ‘Cold Stress’ were more poorly separated (OOB error rate 15.94%). Verifying the RF classification model for the three states ‘Optimal condition’, ‘Cold stress’ and ‘Winter Light Stress’ simultaneously using data from the crown field survey showed that the ‘Winter Light Stress’ state was well identified. In this case, ‘Optimal condition’ was mistakenly defined as ‘Cold stress’. The following vegetation indices were significant for identifying WLS: CARI, CCI, CCRI, CRI₅₅₀, CTRI, LSI, PRI, PRI_m1, modPRI and TVI. Therefore, spectral phenotyping using HSI is a promising method for identifying WLS in conifers. Full article

Satellite-based Earth Observation (EO) has been proposed as an efficient, replicable, and scale-able method for monitoring kelp forests. Although kelps (Laminariales) have been mapped with multispectral EO, no evaluation of kelps’ separability across genera, and from other macroalgae, has been conducted with image-applicable methods. Since kelps and other macroalgae commonly cooccur, characterising their spectral separability is vital to defining appropriate use-cases, methods, and limitations of mapping them with EO. This work investigates the spectral reflectance separability of three kelps and twelve other macroalgae from three distinct regions of Australia and New Zealand. Separability was evaluated using hierarchical clustering, spectral angle, random forest classification, and linear discrimination classification algorithms. Random forest was most effective (average F1 score = 0.70) at classifying all macroalgae by genus, while the linear discriminant analysis was most effective at differentiating among kelp genera labelled by sampling region (average F1 score = 0.93). The observed intra-class geographic variability indicates that macroalgal spectral reflectance is regionally specific, thereby limiting reference spectrum transferability and large-spatial-extent classification accuracy. Of the four classification methods evaluated, the random forest was best suited to mapping large spatial extents (e.g., >100 s km²). Using aggregated target classes is recommended if relying solely on spectral reflectance information. This work suggests hyperspectral EO could be a useful tool in monitoring ecologically and economically valuable kelp forests with moderate to high confidence. Full article

To address the limitations of single indices in comprehensively evaluating the quality of Korla fragrant pears, this study proposes the firmness–soluble solids ratio (FSR), defined as the ratio of average firmness (FI) to soluble solid content (SSC) for each individual fruit, as a novel index. Using 600 samples from five maturity stages with hyperspectral imaging (950–1650 nm), the dataset was split 4:1 by the SPXY algorithm. The findings demonstrated that FSR’s effectiveness in quantifying the dynamic relationship between FI and SSC during maturation. The developed multiscale convolutional neural network–long short-term memory (MSCNN–LSTM) model achieved high prediction accuracy with determination coefficients of 0.8934 (FI), 0.8731 (SSC), and 0.8610 (FSR), and root mean square errors of 0.9001 N, 0.7976%, and 0.1676, respectively. All residual prediction deviation values exceeded 2.5, confirming model robustness. The MSCNN–LSTM showed superior performance compared to other benchmark models. Furthermore, the integration of prediction models with visualization techniques successfully mapped the spatial distribution of quality indices. For maturity discrimination, hyperspectral-based partial least squares discriminant analysis and linear discriminant analysis models achieved perfect classification accuracy (100%) under five-fold cross-validation across all five maturity stages. This work provides both a theoretical basis and a technical framework for non-destructive evaluation of comprehensive quality and maturity in Korla fragrant pears. Full article

Early detection and diagnosis of plant diseases is critical for ensuring global food security and sustainable agricultural practices. This review comprehensively examines latest advancements in crop disease risk prediction, onset detection through imaging techniques, machine learning (ML), deep learning (DL), and edge computing technologies. Traditional disease detection methods, which rely on visual inspections, are time-consuming, and often inaccurate. While chemical analyses are accurate, they can be time consuming and leave less flexibility to promptly implement remedial actions. In contrast, modern techniques such as hyperspectral and multispectral imaging, thermal imaging, and fluorescence imaging, among others can provide non-invasive and highly accurate solutions for identifying plant diseases at early stages. The integration of ML and DL models, including convolutional neural networks (CNNs) and transfer learning, has significantly improved disease classification and severity assessment. Furthermore, edge computing and the Internet of Things (IoT) facilitate real-time disease monitoring by processing and communicating data directly in/from the field, reducing latency and reliance on in-house as well as centralized cloud computing. Despite these advancements, challenges remain in terms of multimodal dataset standardization, integration of individual technologies of sensing, data processing, communication, and decision-making to provide a complete end-to-end solution for practical implementations. In addition, robustness of such technologies in varying field conditions, and affordability has also not been reviewed. To this end, this review paper focuses on broad areas of sensing, computing, and communication systems to outline the transformative potential of end-to-end solutions for effective implementations towards crop disease management in modern agricultural systems. Foundation of this review also highlights critical potential for integrating AI-driven disease detection and predictive models capable of analyzing multimodal data of environmental factors such as temperature and humidity, as well as visible-range and thermal imagery information for early disease diagnosis and timely management. Future research should focus on developing autonomous end-to-end disease monitoring systems that incorporate these technologies, fostering comprehensive precision agriculture and sustainable crop production. Full article

Platycodonis radix is a commonly used traditional Chinese medicine (TCM) material. Its bioactive compounds and medicinal value are closely related to its geographical origin. The internal components of Platycodonis radix from different origins are different due to the influence of environmental factors such as soil and climate. These differences can affect the medicinal value. Therefore, accurate identification of Platycodonis radix origin is crucial for drug safety and scientific research. Traditional methods of identification of TCM materials, such as morphological identification and physicochemical analysis, cannot meet the efficiency requirements. Although emerging technologies such as computer vision and spectroscopy can achieve rapid detection, their accuracy in identifying the origin of Platycodonis radix is limited when relying solely on RGB images or spectral features. To solve this problem, we aim to develop a rapid, non-destructive, and accurate method for origin identification of Platycodonis radix using hyperspectral imaging (HSI) combined with deep learning. We captured hyperspectral images of Platycodonis radix slices in 400–1000 nm range, and proposed a deep learning classification model based on these images. Our model uses one-dimensional (1D) convolution kernels to extract spectral features and two-dimensional (2D) convolution kernels to extract spatial features, fully utilizing the hyperspectral data. The average accuracy has reached 96.2%, significantly better than that of 49.0% based on RGB images and 81.8% based on spectral features in 400–1000 nm range. Furthermore, based on hyperspectral images, our model’s accuracy is 14.6%, 8.4%, and 9.6% higher than the variants of VGG, ResNet, and GoogLeNet, respectively. These results not only demonstrate the advantages of HSI in identifying the origin of Platycodonis radix, but also demonstrate the advantages of combining 1D convolution and 2D convolution in hyperspectral image classification. Full article

Quercetin, a key flavonoid in Anoectochilus roxburghii (Wall.) Lindl., plays an important role in determining the pharmacological value of this medicinal herb. However, traditional methods for quercetin quantification are destructive and time-consuming, limiting their application in real-time quality monitoring. This study investigates the hyperspectral response characteristics of quercetin using near-infrared hyperspectral imaging and establishes a feature-based model to explore its detectability in A. roxburghii leaves. We scanned standard quercetin solutions of known concentration under the same imaging conditions as the leaves to produce a dilution series. Feature-selection methods used included the successive projections algorithm (SPA), Pearson correlation, and competitive adaptive reweighted sampling (CARS). A 1D convolutional neural network (1D-CNN) trained on SPA-selected wavelengths yielded the best prediction performance. These key wavelengths—particularly the 923 nm band—showed strong theoretical and statistical relevance to quercetin’s molecular absorption. When applied to plant leaf spectra, the standard-trained model produced continuous predicted quercetin values that effectively distinguished cultivars with varying flavonoid contents. PCA visualization and ROC-based classification confirmed spectral transferability and potential for functional evaluation. This study demonstrates a non-destructive, spatially resolved, and biochemically interpretable strategy for identifying bioactive markers in plant tissues, offering a methodological basis for future hyperspectral inversion studies and intelligent quality assessment in herbal medicine. Full article

Hyperspectral (HS) remote sensing is a valuable tool for geological surveys and mineral classification. However, mineral maps derived from HS data can exhibit inconsistencies across different imaging times or sensors due to complex factors. In this study, we propose a novel method to enhance the robustness and temporal consistency of mineral mapping. The method combines the spectral identification capabilities of the Tetracorder expert system, developed by United States Geological Survey (USGS), with a data-driven classification model, involving the application of Tetracorder to high-purity pixels identified through the pixel purity index (PPI) analysis to generate reliable training labels. These labels, along with hyperspectral bands transformed by the minimum noise fraction (MNF), are used to train a random forest classifier. The methodology was evaluated using multi-temporal images of the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS), acquired over Cuprite, Nevada, between 2011 and 2013. The results demonstrate that the proposed method achieves accuracy comparable to Tetracorder while improving map consistency and reducing inter-annual mapping errors by approximately 30%. Full article

As climate extremes increasingly threaten global food security, precision tools for early detection of crop stress have become vital, particularly for root crops such as potato (Solanum tuberosum L.) and sweet potato (Ipomoea batatas L. Lam.), which are especially susceptible to environmental stressors throughout their life cycles. In this study, plants were monitored from the initial onset of seasonal stressors, including spring drought, heat, and episodes of excessive rainfall, through to harvest, capturing the full range of physiological and biochemical responses under seasonal, simulated conditions in greenhouses. The spectral data were obtained from regions of interest (ROIs) of each cultivar’s leaves, with over 3000 data points extracted per cultivar; these data were subsequently used for model development. A comprehensive classification framework was established by employing machine learning models, Support Vector Machine (SVM), Linear Discriminant Analysis (LDA), and Partial Least Squares-Discriminant Analysis (PLS-DA), to detect stress across various growth stages. Furthermore, severity levels were objectively defined using photoreflectance indices and principal component analysis (PCA) data visualizations, which enabled consistent and reliable classification of stress responses in both individual cultivars and combined datasets. All models achieved high classification accuracy (90–98%) on independent test sets. The application of the Successive Projections Algorithm (SPA) for variable selection significantly reduced the number of wavelengths required for robust stress classification, with SPA-PLS-DA models maintaining high accuracy (90–96%) using only a subset of informative bands. Furthermore, SPA-PLS-DA-based chemical imaging enabled spatial mapping of stress severity within plant tissues, providing early, non-invasive insights into physiological and biochemical status. These findings highlight the potential of integrating hyperspectral imaging and machine learning for precise, real-time crop monitoring, thereby contributing to sustainable agricultural management and reduced yield losses. Full article

Hyperspectral imaging (HSI) has emerged as a powerful tool for non-destructive phenotyping, yet fruit crop applications remain underexplored. We propose a methodological framework to enhance the spectral characterization of apple fruits by identifying robust vegetation indices (VIs) and interpretable endmembers. We screened 284 Vis, which were evaluated using four feature selection algorithms (Boruta, MI+Lasso, RFE, and ensemble voting), generalizing across red, yellow, green, and purple apple cultivars. An ensemble criterion (≥2 algorithms) yielded 50 selected VIs from the NDSI/DSI/RSI families, preserving > 95% classification accuracy and capturing cultivar-specific variation. Pigment-sensitive wavelength bands were identified via PLS-DA VIP scores and one-vs-rest ANOVA. Using these bands, we formulated a new normalized-difference, ratio, and difference spectral indices tailored to cultivar-specific pigmentation. Several indices achieved >89% classification accuracy and showed patterns consistent with those of anthocyanin, carotenoid, and chlorophyll. A two-stage spectral unmixing pipeline (K-Means → N-FINDR) achieved the lowest reconstruction RMSE (0.043%). This multi-level strategy provides a scalable, interpretable framework for enhancing phenotypic resolution in apple hyperspectral data, contributing to fruit index development and generalized spectral analysis methods for horticultural applications. Full article