Challenging Classification Research Articles

Immunohistochemistry Markers in Ovarian and Fallopian Tube Neoplasms: a Comprehensive Review.

Immunohistochemistry (IHC) has emerged as a crucial tool in diagnosing and managing ovarian cancer, offering invaluable insights into tumor biology and guiding therapeutic decisions. The intricate histopathological landscape of ovarian cancer presents challenges in accurate diagnosis and classification. IHC offers a complementary approach, aiding in the characterization of tumor subtypes, prognostication, and prediction of treatment response. By targeting specific biomarkers, IHC enables the identification of diverse histological features and molecular alterations associated with ovarian malignancies. The integration of IHC into routine diagnostic workflows enhances diagnostic accuracy, aids in the subclassification of ovarian tumors, and facilitates personalized treatment strategies. Emphasis is placed on the judicious selection of antibody panels tailored to specific clinical scenarios, ensuring optimal utilization of resources and minimizing diagnostic pitfalls. Overall, this review underscores the pivotal role of IHC in refining the diagnosis, prognostication, and management of ovarian cancer, highlighting its significance in the era of precision medicine. By leveraging the molecular insights provided by IHC, clinicians and pathologists can optimize patient care and improve outcomes in ovarian cancer management.

Enhancing Image Classification Using Few-Shot Learning Prototypical Networks with ResNet-18: Detection, Accuracy Enhancement, and Optimization

Image classification, especially in scenarios with limited data, presents significant challenges. Few shot learning (FSL) aims to address these challenges by training models that can generalize from a few examples. This paper explores the integration of prototypical networks with ResNet-18 for feature extraction to enhance image classification accuracy. Prototypical networks are designed to create a prototype representation for each class, which can then be used to classify new examples based on their distance to these prototypes. By leveraging ResNet-18's powerful feature extraction capabilities, we aim to improve the quality of these prototypes, thereby enhancing classification performance.We propose various methods for accuracy enhancement and optimization, including hyperparameter tuning, regularization techniques, and advanced methods like attention mechanisms and metric learning. Hyperparameter tuning involves adjusting the model's parameters to find the optimal settings that yield the best performance. Regularization techniques, such as dropout and weight decay, help prevent overfitting and improve the model's generalization capabilities. Advanced methods like attention mechanisms can focus on the most relevant parts of the image, while metric learning aims to learn a distance metric that better reflects the similarities between images.Our experiments on datasets like Mini-ImageNet and Omniglot demonstrate significant improvements in classification performance. These datasets are commonly used benchmarks in the few-shot learning community, allowing us to compare our results with existing methods. The integration of prototypical networks with ResNet-18, along with the proposed optimization techniques, provides a robust approach for tackling the challenges of image classification in few-shot learning scenarios. Key Words: Few-shot learning, ResNet-18, Prototypical Networks.

Categorization challenges in research on organ donation after euthanasia: Determining somatic or psychiatric origins of suffering

This perspective article discusses the classification challenges encountered in organ donation after euthanasia (ODE) among individuals with psychiatric disorders, as highlighted in our recent case series. In cases involving “somatic symptoms and related disorders,” somatic symptoms are connected to excessive thoughts, feelings, and behaviors, causing distress or dysfunction. Conditions such as tinnitus and chronic pain often co-occur with psychiatric disorders. The recent Dutch guideline on ODE eliminated the requirement to distinguish between somatic and psychiatric causes of suffering. This updated guideline encourages a holistic approach to patient care, emphasizing the importance of addressing patients’ overall suffering rather than specific categorizations. To avoid errors in classification, categorization in research on ODE should recognize the overlap between somatic and psychiatric diagnoses, thereby eliminating the risk of misclassification, avoiding the stigmatization of patients, and optimizing future treatment options.

Attention Mechanism-Based Glaucoma Classification Model Using Retinal Fundus Images.

This paper presents a classification model for eye diseases utilizing attention mechanisms to learn features from fundus images and structures. The study focuses on diagnosing glaucoma by extracting retinal vessels and the optic disc from fundus images using a ResU-Net-based segmentation model and Hough Circle Transform, respectively. The extracted structures and preprocessed images were inputted into a CNN-based multi-input model for training. Comparative evaluations demonstrated that our model outperformed other research models in classifying glaucoma, even with a smaller dataset. Ablation studies confirmed that using attention mechanisms to learn fundus structures significantly enhanced performance. The study also highlighted the challenges in normal case classification due to potential feature degradation during structure extraction. Future research will focus on incorporating additional fundus structures such as the macula, refining extraction algorithms, and expanding the types of classified eye diseases.

ODNet: A High Real-Time Network Using Orthogonal Decomposition for Few-Shot Strip Steel Surface Defect Classification.

Strip steel plays a crucial role in modern industrial production, where enhancing the accuracy and real-time capabilities of surface defect classification is essential. However, acquiring and annotating defect samples for training deep learning models are challenging, further complicated by the presence of redundant information in these samples. These issues hinder the classification of strip steel surface defects. To address these challenges, this paper introduces a high real-time network, ODNet (Orthogonal Decomposition Network), designed for few-shot strip steel surface defect classification. ODNet utilizes ResNet as its backbone and incorporates orthogonal decomposition technology to reduce the feature redundancies. Furthermore, it integrates skip connection to preserve essential correlation information in the samples, preventing excessive elimination. The model optimizes the parameter efficiency by employing Euclidean distance as the classifier. The orthogonal decomposition not only helps reduce redundant image information but also ensures compatibility with the Euclidean distance requirement for orthogonal input. Extensive experiments conducted on the FSC-20 benchmark demonstrate that ODNet achieves superior real-time performance, accuracy, and generalization compared to alternative methods, effectively addressing the challenges of few-shot strip steel surface defect classification.

Lung Sound Classification for Respiratory Disease Identification Using Deep Learning: A Survey

Integrating artificial intelligence (AI) into lung sound classification has markedly improved respiratory disease diagnosis by analysing intricate patterns within audio data. This study is driven by the widespread issue of lung diseases, which affect around 500 million people globally. Early detection of respiratory diseases is crucial for delivering timely and effective treatment. Our study consists of a comprehensive survey of lung sound classification methodologies, exploring the advancements made in leveraging AI to identify and classify respiratory diseases. This survey thoroughly investigates lung sound classification models, along with data augmentation, feature extraction, explainable techniques and support tools to improve systems for diagnosing respiratory conditions. Our goal is to provide meaningful insights for healthcare professionals, researchers and technologists who are dedicated to developing methodologies for the early detection of pulmonary diseases. The paper provides a summary of the current status of lung sound classification research, highlighting both advancements and challenges in the use of AI for more accurate and efficient diagnostic methods in respiratory healthcare.

Skin lesion classification using modified deep and multi-directional invariant handcrafted features

Skin lesions encompass various skin conditions, including cancerous growths resulting from uncontrolled proliferation of skin cells. Globally, this disease affects a significant portion of the population, with millions of fatalities recorded. Over the past three decades, there has been a concerning escalation in diagnosed cases of skin cancer. Early detection is crucial for effective treatment, as late diagnosis significantly heightens mortality risk. Existing research often focuses on either handcrafted or deep features, neglecting the diverse textural and structural properties inherent in skin lesion images. Additionally, reliance on a single optimizer in CNN-based schemes poses efficiency challenges. To tackle these issues, this paper presents two novel approaches for classifying skin lesions in dermoscopic images to assess cancer severity. The first approach enhances classification accuracy by leveraging a modified VGG-16 network and employing both RMSProp and Adam optimizers. The second approach introduces a Hybrid CNN Model, integrating deep features from the modified VGG-16 network with handcrafted color and multi-directional texture features. Color features are extracted using a non-uniform cumulative probability-based histogram method, while texture features are derived from a 45∘ rotated complex wavelet filter-based dual-tree complex wavelet transform. The amalgamated features facilitate accurate prediction of skin lesion classes. Evaluation on ISIC 2017 skin cancer classification challenge images demonstrates significant performance enhancements over existing techniques.

RNDDNet: A residual nested dilated DenseNet based deep-learning model for chilli plant disease classification

The most significant peril to food safety arises from plant diseases, capable of substantially diminishing both the quantity and quality of agricultural yields. Identifying these plant diseases stands out as the foremost challenge within the agricultural sector. Convolutional and deep neural networks prove effective in resolving image classification challenges within the realm of computer vision. Numerous Deep Neural Network(DNN)-based structures have been employed to diagnose plant diseases. Many DNN models in the field make use of various iterations of Dense and DenseNet layers in order to enhance the receptive field and capture intricate features within the data. However, it is important to note that such models often come with a significant computational burden and can introduce aliasing artifacts due to their complexity and resource-intensive nature. To overcome those limitations, we proposed a novel Residual Nested Dilated DenseNet based deep-learning (RNDDNet) model in this paper. Residual Nested Dilated DenseNet model residual connections are achieving the required receptive field, and their dilation factors are effective in extracting more features. The RNDDNet model exhibits the highest level of accuracy in identifying plant diseases. This research introduces a less computational cost and compact model to detect diseases in plant leaves. The proposed model functions to identify diseases, utilizing a dataset comprising 3,800 photographs of chilli leaves, categorized into six distinct classes: five disorder classes and one healthy chilli class. Through experimentation, the outcomes indicate that the suggested model achieves an accuracy of 98.09 %, along with a precision of 97 %, a recall of 97.25 %, and an F1 score of 97.25%. The presented approach demonstrates its superiority over existing methodologies.

Systematic Analysis of Deep Learning Models vs. Machine Learning

Deep learning (DL) and classical machine learning (ML) models are compared and contrasted in this study, which offers a complete overview of the differences and technological improvements between the two types of models. Through an analysis of a diverse range of research publications, the study draws attention to the distinct advantages and uses of both techniques. Deep learning, which is characterized by its use of neural networks with several layers, is particularly effective at managing massive datasets that are not organized. It has also made great progress in the areas of image and audio recognition, natural language processing, and complicated pattern identification exercises. On the other hand, classic machine learning models, which are based on the extraction of features and simpler methods, continue to be quite successful in structured data situations such as classification, regression, and clustering challenges. By concentrating on aspects such as data quantity, computing resources, and unique application needs, the survey sheds light on the parameters that should be considered when selecting between deep learning and machine learning. In addition to this, it addresses the ever-changing environment of hybrid models, which combine methods from both deep learning and machine learning in order to capitalize on the advantages of both approaches. This study highlights the significance of contextual awareness in the fast-developing area of artificial intelligence by providing researchers and practitioners with useful insights that can be used to deploy the AI models that are the most appropriate for their particular requirements.

O-110 Adenomyosis : the challenge of classification and diagnosis

Abstract Study question What does the collaborative document on adenomyosis terminology, classification, and diagnosis by ESGE, ESHRE, and WES provides? Summary answer This good practice recommendations paper offers guidance on the technical aspects, terminology, classification, and diagnosis of adenomyosis in women of reproductive age. What is known already The existing literature on adenomyosis terminology is heterogeneous and confusing. Various classification systems based on ultrasound and MRI have been proposed, yet none have gained international acceptance or clinical validation. The contradictory data on the diagnostic accuracy of these modalities further complicate adenomyosis diagnosis in reproductive-age women. Study design, size, duration A working group from ESGE, ESHRE, and WES has collaborated to write recommendations concerning the terminology, classification, and diagnosis of adenomyosis. Participants/materials, setting, methods Focusing on terminology, classification, and diagnosis, this good practice paper paves the way for further recommendations on the management and surgical treatment of adenomyosis in reproductive-age women. Main results and the role of chance The document overviews adenomyosis terminology and reviews proposed classification systems based on histology, ultrasound, and MRI, concluding with diagnostic recommendations. Limitations, reason for caution Owing the scarcity of evidence, these recommendations predominantly rely on clinical expertise. Wider implications of the findings The recommendations aim to enhance existing adenomyosis guidelines. Study funding/Competing interests All costs were covered by ESHRE funds, with no external funding for the development process or manuscript production.

Widened Attention‐Enhanced Atrous Convolutional Network for Efficient Embedded Vision Applications under Resource Constraints

Onboard image analysis enables real‐time autonomous capabilities for unmanned platforms including aerial, ground, and aquatic drones. Performing classification on embedded systems, rather than transmitting data, allows rapid perception and decision‐making critical for time‐sensitive applications such as search and rescue, hazardous environment exploration, and military operations. To fully capitalize on these systems’ potential, specialized deep learning solutions are needed that balance accuracy and computational efficiency for time‐sensitive inference. This article introduces the widened attention‐enhanced atrous convolution‐based efficient network (WACEfNet), a new convolutional neural network designed specifically for real‐time visual classification challenges using resource‐constrained embedded devices. WACEfNet builds on EfficientNet and integrates innovative width‐wise feature processing, atrous convolutions, and attention modules to improve representational power without excessive overhead. Extensive benchmarking confirms state‐of‐the‐art performance from WACEfNet for aerial imaging applications while remaining suitable for embedded deployment. The improvements in accuracy and speed demonstrate the potential of customized deep learning advancements to unlock new capabilities for unmanned aerial vehicles and related embedded systems with tight size, weight, and power constraints. This research offers an optimized framework, combining widened residual learning and attention mechanisms, to meet the unique demands of high‐fidelity real‐time analytics across a variety of embedded perception paradigms.

COVIDHealth: A novel labeled dataset and machine learning-based web application for classifying COVID-19 discourses on Twitter

The COVID-19 pandemic has sparked widespread health-related discussions on social media platforms like Twitter (now named ‘X’). However, the lack of labeled Twitter data poses significant challenges for theme-based classification and tweet aggregation. To address this gap, we developed a machine learning-based web application that automatically classifies COVID-19 discourses into five categories: health risks, prevention, symptoms, transmission, and treatment. We collected and labeled 6,667 COVID-19-related tweets using the Twitter API, and applied various feature extraction methods to extract relevant features. We then compared the performance of seven classical machine learning algorithms (Decision Tree, Random Forest, Stochastic Gradient Descent, Adaboost, K-Nearest Neighbor, Logistic Regression, and Linear SVC) and four deep learning techniques (LSTM, CNN, RNN, and BERT) for classification. Our results show that the CNN achieved the highest precision (90.41%), recall (90.4%), F1 score (90.4%), and accuracy (90.4%). The Linear SVC algorithm exhibited the highest precision (85.71%), recall (86.94%), and F1 score (86.13%) among classical machine learning approaches. Our study advances the field of health-related data analysis and classification, and offers a publicly accessible web-based tool for public health researchers and practitioners. This tool has the potential to support addressing public health challenges and enhancing awareness during pandemics. The dataset and application are accessible at https://github.com/Bishal16/COVID19-Health-Related-Data-Classification-Website.

Enhancing ECG Heartbeat classification with feature fusion neural networks and dynamic minority-biased batch weighting loss function

Objective. This study aims to address the challenges of imbalanced heartbeat classification using electrocardiogram (ECG). In this proposed novel deep-learning method, the focus is on accurately identifying minority classes in conditions characterized by significant imbalances in ECG data. Approach. We propose a feature fusion neural network enhanced by a dynamic minority-biased batch weighting loss function. This network comprises three specialized branches: the complete ECG data branch for a comprehensive view of ECG signals, the local QRS wave branch for detailed features of the QRS complex, and the R wave information branch to analyze R wave characteristics. This structure is designed to extract diverse aspects of ECG data. The dynamic loss function prioritizes minority classes while maintaining the recognition of majority classes, adjusting the network’s learning focus without altering the original data distribution. Together, this fusion structure and adaptive loss function significantly improve the network’s ability to distinguish between various heartbeat classes, enhancing the accuracy of minority class identification. Main results. The proposed method demonstrated balanced performance within the MIT-BIH dataset, especially for minority classes. Under the intra-patient paradigm, the accuracy, sensitivity, specificity, and positive predictive value for Supraventricular ectopic beat were 99.63 % , 93.62 % , 99.81 % , and 92.98 % , respectively, and for Fusion beat were 99.76 % , 85.56 % , 99.87 % , and 84.16 % , respectively. Under the inter-patient paradigm, these metrics were 96.56 % , 89.16 % , 96.84 % , and 51.99 % for Supraventricular ectopic beat, and 96.10 % , 77.06 % , 96.25 % , and 13.92 % for Fusion beat, respectively. Significance. This method effectively addresses the class imbalance in ECG datasets. By leveraging diverse ECG signal information and a novel loss function, this approach offers a promising tool for aiding in the diagnosis and treatment of cardiac conditions.

Multi-class AUC maximization for imbalanced ordinal multi-stage tropical cyclone intensity change forecast

Intense tropical cyclones (TCs) cause significant damage to human societies. Forecasting the multiple stages of TC intensity changes is considerably crucial yet challenging. This difficulty arises due to imbalanced data distribution and the need for ordinal multi-class classification. While existing classification methods, such as linear discriminant analysis, have been utilized to predict rare rapidly intensifying (RI) stages based on features related TC intensity changes, they are limited to binary classification distinguishing between RI and non-RI stages. In this paper, we introduce a novel methodology to tackle the challenges of imbalanced ordinal multi-class classification. We extend the Area Under the Curve maximization technique with inter-instance/class cross-hinge losses and inter-class distance-based slack variables. The proposed loss function, implemented within a deep learning framework, demonstrates its effectiveness using real sequence data of multi-stage TC intensity changes, including satellite infrared images and environmental variables observed in the western North Pacific.

Molecular delimitation of cryptic Australian squid species of the genus Uroteuthis Rehder, 1945 (Cephalopoda: Loliginidae), provides a baseline of diversity to resolve classification challenges throughout the Indo-Pacific

ABSTRACT This study provides a comprehensive molecular phylogenetic analysis of Uroteuthis squid from the Indo-Pacific region. The main aim was to increase sample coverage from northern and eastern Australian waters to resolve the identity and distribution of Uroteuthis species taken by local fisheries. Two mitochondrial regions, cytochrome c oxidase subunit I (COI) and 16S ribosomal RNA (16S rRNA), were sequenced from 220 new specimens and analysed with a further 51 sequences from GenBank to create a combined phylogeny for the genus. Three nuclear regions, 18S ribosomal DNA (18S rDNA), 28S ribosomal DNA (28S rDNA) and rhodopsin, were also sequenced from representatives of each species. Based on the mitochondrial phylogeny plus distance and tree-based delimitation models, a COI species barcode gap of 4–5% is proposed for discriminating Uroteuthis species. Applying this gap partitioned many described species into species complexes; for example, U. duvaucelii, U. noctiluca and U. edulis resolved into 10 species. Although more conserved, mitochondrial 16S rRNA sequences differentiated all new species clades, whereas none of the nuclear markers resolved the closest species. Results confirm that neither U. chinensis nor U. edulis occurs in Australian waters. Five undescribed species are identified from northern and eastern Australia, of which four are consistent with earlier allozyme studies (and two align with existing DNA sequences). One is a new southeastern, deeper shelf species differentiated in this study, along with a sixth undescribed species from Indonesian waters. Results of the molecular analysis are now being used to inform complementary morphometric analyses for new species descriptions, and genetic stock structure assessments of these important fisheries resources.

Fluorescence excitation-scanning hyperspectral imaging with scalable 2D–3D deep learning framework for colorectal cancer detection

Colorectal cancer is one of the top contributors to cancer-related deaths in the United States, with over 100,000 estimated cases in 2020 and over 50,000 deaths. The most common screening technique is minimally invasive colonoscopy using either reflected white light endoscopy or narrow-band imaging. However, current imaging modalities have only moderate sensitivity and specificity for lesion detection. We have developed a novel fluorescence excitation-scanning hyperspectral imaging (HSI) approach to sample image and spectroscopic data simultaneously on microscope and endoscope platforms for enhanced diagnostic potential. Unfortunately, fluorescence excitation-scanning HSI datasets pose major challenges for data processing, interpretability, and classification due to their high dimensionality. Here, we present an end-to-end scalable Artificial Intelligence (AI) framework built for classification of excitation-scanning HSI microscopy data that provides accurate image classification and interpretability of the AI decision-making process. The developed AI framework is able to perform real-time HSI classification with different speed/classification performance trade-offs by tailoring the dimensionality of the dataset, supporting different dimensions of deep learning models, and varying the architecture of deep learning models. We have also incorporated tools to visualize the exact location of the lesion detected by the AI decision-making process and to provide heatmap-based pixel-by-pixel interpretability. In addition, our deep learning framework provides wavelength-dependent impact as a heatmap, which allows visualization of the contributions of HSI wavelength bands during the AI decision-making process. This framework is well-suited for HSI microscope and endoscope platforms, where real-time analysis and visualization of classification results are required by clinicians.

Augmenting Telepostpartum Care With Vision-Based Detection of Breastfeeding-Related Conditions: Algorithm Development and Validation.

Breastfeeding benefits both the mother and infant and is a topic of attention in public health. After childbirth, untreated medical conditions or lack of support lead many mothers to discontinue breastfeeding. For instance, nipple damage and mastitis affect 80% and 20% of US mothers, respectively. Lactation consultants (LCs) help mothers with breastfeeding, providing in-person, remote, and hybrid lactation support. LCs guide, encourage, and find ways for mothers to have a better experience breastfeeding. Current telehealth services help mothers seek LCs for breastfeeding support, where images help them identify and address many issues. Due to the disproportional ratio of LCs and mothers in need, these professionals are often overloaded and burned out. This study aims to investigate the effectiveness of 5 distinct convolutional neural networks in detecting healthy lactating breasts and 6 breastfeeding-related issues by only using red, green, and blue images. Our goal was to assess the applicability of this algorithm as an auxiliary resource for LCs to identify painful breast conditions quickly, better manage their patients through triage, respond promptly to patient needs, and enhance the overall experience and care for breastfeeding mothers. We evaluated the potential for 5 classification models to detect breastfeeding-related conditions using 1078 breast and nipple images gathered from web-based and physical educational resources. We used the convolutional neural networks Resnet50, Visual Geometry Group model with 16 layers (VGG16), InceptionV3, EfficientNetV2, and DenseNet169 to classify the images across 7 classes: healthy, abscess, mastitis, nipple blebs, dermatosis, engorgement, and nipple damage by improper feeding or misuse of breast pumps. We also evaluated the models' ability to distinguish between healthy and unhealthy images. We present an analysis of the classification challenges, identifying image traits that may confound the detection model. The best model achieves an average area under the receiver operating characteristic curve of 0.93 for all conditions after data augmentation for multiclass classification. For binary classification, we achieved, with the best model, an average area under the curve of 0.96 for all conditions after data augmentation. Several factors contributed to the misclassification of images, including similar visual features in the conditions that precede other conditions (such as the mastitis spectrum disorder), partially covered breasts or nipples, and images depicting multiple conditions in the same breast. This vision-based automated detection technique offers an opportunity to enhance postpartum care for mothers and can potentially help alleviate the workload of LCs by expediting decision-making processes.

Measles and Rubella Diagnostic and Classification Challenges in Near- and Post-Elimination Countries.

Measles and rubella are vaccine-preventable viral diseases and can be prevented by safe, highly effective vaccination with measles- and rubella-containing vaccines. Given the myriad causes of febrile exanthems, laboratory surveillance for both measles and rubella is important to document the incidence of these diseases and to track the progress and maintenance of elimination in near- and post-elimination settings. Diagnostic challenges can hinder effective surveillance and classification challenges can hinder efforts to demonstrate achievement or maintenance of elimination. In this report, we review diagnostic and classification challenges for measles and rubella in near- and post-elimination settings.

SRFAD-Net: Scale-Robust Feature Aggregation and Diffusion Network for Object Detection in Remote Sensing Images

The significant differences in target scales of remote sensing images lead to remarkable variations in visual features, posing significant challenges for feature extraction, fusion, regression, and classification. For example, models frequently struggle to capture features of targets across all scales, inadequately consider the weights and importance of features at different scales during fusion, and encounter accuracy limitations when detecting targets of varying scales. To tackle these challenges, we proposes a Scale-Robust Feature Aggregation and Diffusion Network (SRFAD-Net) for remote sensing target detection. This model includes a Scale-Robust Feature Network (SRFN), an Adaptive Feature Aggregation and Diffusion (AFAD) module, and a Focaler-GIoU Loss. SRFN extracts scale-robust features by constructing a multi-scale pyramid. It includes a downsampling (ADown) module that combines the advantages of average pooling and max pooling, effectively preserving background information and salient features. This further enhances the network’s ability to handle targets of varying scales and shapes. The introduced Deformable Attention(DAttention) mechanism captures target features effectively through adaptive adjustment of the receptive field’s shape and size, reducing background clutter and substantially enhancing the model’s performance in detecting distant objects. In the feature fusion stage, we propose the AFAD module, which utilizes a dimension-adaptive perceptual selection mechanism and parallel depthwise convolutional operations to precisely aggregate multi-channel information. It then employs a diffusion mechanism to spread contextual information across various scales, greatly improving the network’s ability to extract and fuse features across multiple scales. For the detection head, we adopt the Focaler-GIoU Loss, leveraging its advantages in handling non-overlapping bounding boxes, effectively alleviating the difficulty of localization caused by scale variations. We have undertaken experiments on two widely utilized aerial target datasets: the Remote Sensing Scene Object Detection Dataset (RSOD) and NWPU VHR-10, which is a high-resolution object detection dataset from Northwestern Polytechnical University. The findings of these experiments clearly illustrate that SRFAD-Net surpasses the performances of mainstream detectors.

Cross-modal and multimodal data analysis based on functional mapping of spectral descriptors and manifold regularization

Multimodal manifold modeling methods extend spectral geometry-aware data analysis to incorporate learning from multiple related and complementary modalities. However, most methods assume an equal number of homogeneous data samples in each modality and partial correspondences between modalities as prior knowledge. This work introduces two new multimodal modeling methods. The first method introduces a comprehensive framework for handling multimodal information in heterogeneous data without requiring specific prior knowledge. To achieve this, we begin by extracting local descriptors using spectral graph wavelet signatures (SGWS) to identify manifold localities. Then, we propose a manifold regularization framework that incorporates functional mapping between SGWS descriptors (FMBSD) to determine pointwise correspondences. The second method involves manifold regularized multimodal classification based on pointwise correspondences (M2CPC). This method addresses the challenge of multiclass classification in multimodal heterogeneous data by determining correspondences between modalities using the FMBSD method. Experimental results evaluating the FMBSD method on three widely used cross-modal retrieval datasets and evaluating the M2CPC method on three benchmark multimodal multiclass classification datasets demonstrate their effectiveness and superiority over state-of-the-art methods.