Search Results (403)

Accurate synthetic image generation is crucial for addressing data scarcity challenges in medical image classification tasks, particularly in sensor-derived medical imaging. In this work, we propose a novel method using a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) and nearest-neighbor interpolation to generate high-quality synthetic images for diabetic retinopathy classification. Our approach enhances training datasets by generating realistic retinal images that retain critical pathological features. We evaluated the method across multiple retinal image datasets, including Retinal-Lesions, Fine-Grained Annotated Diabetic Retinopathy (FGADR), Indian Diabetic Retinopathy Image Dataset (IDRiD), and the Kaggle Diabetic Retinopathy dataset. The proposed method outperformed traditional generative models, such as conditional GANs and PathoGAN, achieving the best performance on key metrics: a Fréchet Inception Distance (FID) of 15.21, a Mean Squared Error (MSE) of 0.002025, and a Structural Similarity Index (SSIM) of 0.89 in the Kaggle dataset. Additionally, expert evaluations revealed that only 56.66% of synthetic images could be distinguished from real ones, demonstrating the high fidelity and clinical relevance of the generated data. These results highlight the effectiveness of our approach in improving medical image classification by generating realistic and diverse synthetic datasets. Full article

Electroencephalogram (EEG)-based emotion recognition has garnered significant attention in brain–computer interface research and healthcare applications. While deep learning models have been extensively studied, most are designed for classification tasks and struggle to accurately predict continuous emotional scores in regression settings. In this paper, we introduce EEG-RegNet, a novel deep neural network tailored for precise emotional score prediction across the continuous valence–arousal–dominance (VAD) space. EEG-RegNet tackles two core challenges: extracting subject-independent, emotion-relevant EEG features and mapping these features to fine-grained, continuous emotional scores. The model leverages 2D convolutional neural networks (CNNs) for spatial feature extraction and a 1D CNN for temporal dynamics, providing robust spatiotemporal modeling. A key innovation is the hybrid loss function, which integrates mean squared error (MSE) and cross-entropy (CE) with a Bernoulli penalty to enhance probability estimation and address sparsity in the emotional space. Extensive experiments on the DEAP dataset show that EEG-RegNet achieves state-of-the-art results in continuous emotional score prediction and attains 95% accuracy in fine-grained emotion classification, highlighting its scalability and precision in emotion recognition. Full article

Parameter tuning based adapter methods have achieved notable success in vision-language models (VLMs). However, they face challenges in scenarios with insufficient training samples or limited resources. While leveraging image modality caching and retrieval techniques can reduce resource requirements, these approaches often overlook the significance of textual modality and cross-modal cues in VLMs. To address this, we propose a Conditional Cross-Modal learning model, which is abbreviated as CoCM. CoCM builds separate cache models for both the text and image modalities and embedding textual knowledge conditioned on image information. It dynamically adjusts the cross-modal fusion affinity ratio and disentangles similarity measures across different modalities. Additionally, CoCM incorporates intra-batch image similarity loss as a regularization term to identify hard samples and enhance fine-grained classification performance. CoCM surpasses existing methods in terms of accuracy, generalization ability, and efficiency, achieving a 0.28% accuracy improvement over XMAdapter across 11 datasets and demonstrating 44.79% generalization performance on four cross-domain datasets. Full article

Bipartite networks are common in real-world applications, where link prediction helps understand network evolution and make recommendations. Traditional methods have two major limitations; they often ignore edge weight information by treating links as binary connections, and they struggle to capture complex interaction patterns in sparse networks. We propose WePred, a weight-guided contrastive learning method for link prediction which addresses both challenges through three key components: (1) a weight-guided edge attention mechanism that incorporates edge weights into neighbor aggregation using dynamic attention scores, enabling the fine-grained capture of interaction strengths; (2) a dual-level contrastive learning approach that combines edge-level and node-level contrasts to capture both local weighted patterns and global structural dependencies, which is particularly effective in sparse regions; and (3) a unified learning framework that integrates classification and contrastive objectives. We evaluate WePred on five real-world datasets. The experimental results show that WePred consistently outperformed state-of-the-art methods, achieving improvements of 7.27% in the AUC on ML-1M and 3.18% in the AUC on the sparsest dataset of Amazon-Book (density: 0.001%). Ablation studies confirm the effectiveness of each component, while parameter analyses provide deployment guidance. Full article

The quality of silkworm cocoons affects the quality and cost of silk processing. It is necessary to sort silkworm cocoons prior to silk production. Cocoon images consist of fine-grained images with large intra-class differences and small inter-class differences. The subtle intra-class features pose a serious challenge in accurately locating the effective areas and classifying silkworm cocoons. To improve the perception of intra-class features and the classification accuracy, this paper proposes a bilinear pooling classification model (B-Res41-ASE) based on adaptive multi-scale feature fusion and enhancement. B-Res41-ASE consists of three parts: a feature extraction module, a feature fusion module, and a feature enhancement module. Firstly, the backbone network, ResNet41, is constructed based on the bilinear pooling algorithm to extract complete cocoon features. Secondly, the adaptive spatial feature fusion module (ASFF) is introduced to fuse different semantic information to solve the problem of fine-grained information loss in the process of feature extraction. Finally, the squeeze and excitation module (SE) is used to suppress redundant information, enhance the weight of distinguishable regions, and reduce classification bias. Compared with the widely used classification network, the proposed model achieves the highest classification performance in the test set, with accuracy of 97.0% and an F1-score of 97.5%. The accuracy of B-Res41-ASE is 3.1% and 2.6% higher than that of the classification networks AlexNet and GoogLeNet, respectively, while the F1-score is 2.5% and 2.2% higher, respectively. Additionally, the accuracy of B-Res41-ASE is 1.9% and 7.7% higher than that of the Bilinear CNN and HBP, respectively, while the F1-score is 1.6% and 5.7% higher. The experimental results show that the proposed classification model without complex labelling outperforms other cocoon classification algorithms in terms of classification accuracy and robustness, providing a theoretical basis for the intelligent sorting of silkworm cocoons. Full article

Although significant progress has been made in sentiment analysis tasks based on image–text data, existing methods still have limitations in capturing cross-modal correlations and detailed information. To address these issues, we propose a Multi-Granularity Attention Fusion Network for Implicit Sentiment Analysis (MGAFN-ISA). MGAFN-ISA that leverages neural networks and attention mechanisms to effectively reduce noise interference between different modalities and captures distinct, fine-grained visual and textual features. The model includes two key feature extraction modules: a multi-scale attention fusion-based visual feature extractor and a hierarchical attention mechanism-based textual feature extractor, each designed to extract detailed and discriminative visual and textual representations. Additionally, we introduce an image translator engine to produce accurate and detailed image descriptions, further narrowing the semantic gap between the visual and textual modalities. A bidirectional cross-attention mechanism is also incorporated to utilize correlations between fine-grained local regions across modalities, extracting complementary information from heterogeneous visual and textual data. Finally, we designed an adaptive multimodal classification module that dynamically adjusts the contribution of each modality through an adaptive gating mechanism. Extensive experimental results demonstrate that MGAFN-ISA achieves a significant performance improvement over nine state-of-the-art methods across multiple public datasets, validating the effectiveness and advancement of our proposed approach. Full article

Background/Objectives: Vision Transformers (ViTs) and convolutional neural networks (CNNs) have demonstrated remarkable performances in image classification, especially in the domain of medical imaging analysis. However, ViTs struggle to capture high-frequency components of images, which are critical in identifying fine-grained patterns, while CNNs have difficulties in capturing long-range dependencies due to their local receptive fields, which makes it difficult to fully capture the spatial relationship across lung regions. Methods: In this paper, we proposed a hybrid architecture that integrates ViTs and CNNs within a modular component block(s) to leverage both local feature extraction and global context capture. In each component block, the CNN is used to extract the local features, which are then passed through the ViT to capture the global dependencies. We implemented a gated attention mechanism that combines the channel-, spatial-, and element-wise attention to selectively emphasize the important features, thereby enhancing overall feature representation. Furthermore, we incorporated a multi-scale fusion module (MSFM) in the proposed framework to fuse the features at different scales for more comprehensive feature representation. Results: Our proposed model achieved an accuracy of 99.50% in the classification of four pulmonary conditions. Conclusions: Through extensive experiments and ablation studies, we demonstrated the effectiveness of our approach in improving the medical image classification performance, while achieving good calibration results. This hybrid approach offers a promising framework for reliable and accurate disease diagnosis in medical imaging. Full article

Traditional gross ecosystem product (GEP) accounting methods often operate at macro scales, failing to reflect the localized and nuanced values of wetland ecosystems. This study addresses these challenges by introducing a fine-grained classification system based on a localized adaptation of international standards. The framework integrates high-precision national land surveys and remote sensing quantitative analysis while incorporating fisheries resource models, climate regulation beneficiary mapping, and visitor interpolation to address data scarcity related to human activities. This approach refines the spatial calculation methods for functional quantity accounting at fine scales. The results demonstrate that the refined classification maintains consistency with traditional methods in total value while adapting to multi-scale accounting, filling gaps at small and medium scales and providing a more accurate representation of localized wetland characteristics. Additionally, the study highlights the dominance of cultural services in GEP, emphasizing the need to balance cultural and regulatory services to ensure fairness in decision-making. Finally, a village-scale decision-support model is proposed, offering actionable guidance for wetland management and sustainable development planning. Full article

The assessment of grain size and sediment output is crucial for analyzing the pace of sediment erosion, engineering dams and reservoirs, anticipating the impact of climate change and human activities on river systems, and comprehending the presence of trace and heavy metal pathogens and micropollutants. In July 2024, 16 samples of bed sediments were collected from the mainstream of Wadi Fatima and its tributaries in the Makkah region of Saudi Arabia to identify the depositional environments and the hydrodynamic conditions using Passega diagram, Linear Discriminate Function (LDF) and bivariate plots. The results indicate that the sediments being studied exhibit polymodal properties in both the upstream and midstream regions of the main channel of Wadi Fatima. However, in the downstream region, the samples show trimodal properties. Regarding tributaries, the WFT1, WFT2, and WFT4 sediments exhibit polymodal properties, except for WFT3, which is bimodal. Folk’s classification system categorizes the samples into four distinct classes/facies: gravel, sandy gravel, gravely sand, and sand, with respective proportions of 13%, 62%, 6%, and 19%. The sediments found in Wadi Fatima contain a range of graphic mean (M_Z) values, from −3.34 (indicating medium gravel) to 2.48 (indicating fine sand). On average, the M_Z value is −0.79, which shows extremely fine gravel. The standard deviation (sorting (σi)) values of the samples analyzed from Wadi Fatima vary between 0.71 (moderately sorted) and 3.44 (very poorly sorted), with an average of 2.00 (very poorly sorted). The data exhibits a range of skewness (Sk) values, ranging from −0.41 (showing a very coarse Sk) to 0.82 (representing a indicating a very fine Sk). On average, the data shows a Sk value of −0.02, indicating a symmetrical distribution. The kurtosis (K) values span from 0.51 (indicating a very platykurtic distribution) to 2.65 (indicating a very leptokurtic distribution), with an average of 0.95 (indicating a mesokurtic distribution). Full article

Anxiety is a widespread mental health issue, and binaural beats have been explored as a potential non-invasive treatment. EEG data reveal changes in neural oscillation and connectivity linked to anxiety reduction; however, harmonics introduced during signal acquisition and processing often distort these findings. Existing methods struggle to effectively reduce harmonics and capture the fine-grained temporal dynamics of EEG signals, leading to inaccurate feature extraction. Hence, a novel Denoised Harmonic Subtraction and Transient Temporal Feature Extraction is proposed to improve the analysis of the impact of binaural beats on anxiety levels. Initially, a novel Wiener Fused Convo Filter is introduced to capture spatial features and eliminate linear noise in EEG signals. Next, an Intrinsic Harmonic Subtraction Network is employed, utilizing the Attentive Weighted Least Mean Square (AW-LMS) algorithm to capture nonlinear summation and resonant coupling effects, effectively eliminating the misinterpretation of brain rhythms. To address the challenge of fine-grained temporal dynamics, an Embedded Transfo XL Recurrent Network is introduced to detect and extract relevant parameters associated with transient events in EEG data. Finally, EEG data undergo harmonic reduction and temporal feature extraction before classification with a cross-correlated Markov Deep Q-Network (DQN). This facilitates anxiety level classification into normal, mild, moderate, and severe categories. The model demonstrated a high accuracy of 95.6%, precision of 90%, sensitivity of 93.2%, and specificity of 96% in classifying anxiety levels, outperforming previous models. This integrated approach enhances EEG signal processing, enabling reliable anxiety classification and offering valuable insights for therapeutic interventions. Full article

In this study, we propose a one-branch IncepTAE network to extract local and global hybrid temporal attention simultaneously and congruously for fine-grained satellite image time series (SITS) classification. Transformer and the temporal self-attention mechanism have been the research focus of SITS classification in recent years. However, its effectiveness seems to diminish in the scenario of fine-grained classification among similar categories, for example, different crop types. Theoretically, most of the existing methods focus on only one type of temporal attention, either global attention or local attention, but actually, both of them are required to achieve fine-grained classification. Even though some works adopt two-branch architecture to extract hybrid attention, they usually lack congruity between different types of temporal attention and hinder the expected discriminating ability. Compared with the existing methods, IncepTAE exhibits multiple methodological novelties. Firstly, we insert average/maximum pooling layers into the calculation of multi-head attention to extract hybrid temporal attention. Secondly, IncepTAE adopts one-branch architecture, which reinforces the interaction and congruity of different temporal information. Thirdly, the proposed IncepTAE is more lightweight due to the use of group convolutions. IncepTAE achieves 95.65% and 97.84% overall accuracy on two challenging datasets, TimeSen2Crop and Ghana. The comparative results with existing state-of-the-art methods demonstrate that IncepTAE is able to achieve superior classification performance and faster inference speed, which is conducive to the large-area application of SITS classification. Full article

Pesticide registration information is an essential part of the pesticide knowledge base. However, the large amount of unstructured text data that it contains pose significant challenges for knowledge storage, retrieval, and utilization. To address the characteristics of pesticide registration text such as high information density, complex logical structures, large spans between entities, and heterogeneous entity lengths, as well as to overcome the challenges faced when using traditional joint extraction methods, including triplet overlap, exposure bias, and redundant computation, we propose a single-stage entity–relation joint extraction model based on HT-BES multi-dimensional labeling (MD-SERel). First, in the encoding layer, to address the complex structural characteristics of pesticide registration texts, we employ RoBERTa combined with a multi-head self-attention mechanism to capture the deep semantic features of the text. Simultaneously, syntactic features are extracted using a syntactic dependency tree and graph neural networks to enhance the model’s understanding of text structure. Subsequently, we integrate semantic and syntactic features, enriching the character vector representations and thus improving the model’s ability to represent complex textual data. Secondly, in the multi-dimensional labeling framework layer, we use HT-BES multi-dimensional labeling, where the model assigns multiple labels to each character. These labels include entity boundaries, positions, and head–tail entity association information, which naturally resolves overlapping triplets. Through utilizing a parallel scoring function and fine-grained classification components, the joint extraction of entities and relations is transformed into a multi-label sequence labeling task based on relation dimensions. This process does not involve interdependent steps, thus enabling single-stage parallel labeling, preventing exposure bias and reducing computational redundancy. Finally, in the decoding layer, entity–relation triplets are decoded based on the predicted labels from the fine-grained classification. The experimental results demonstrate that the MD-SERel model performs well on both the Pesticide Registration Dataset (PRD) and the general DuIE dataset. On the PRD, compared to the optimal baseline model, the training time is 1.2 times faster, the inference time is 1.2 times faster, and the F1 score is improved by 1.5%, demonstrating its knowledge extraction capabilities in pesticide registration documents. On the DuIE dataset, the MD-SERel model also achieved better results compared to the baseline, demonstrating its strong generalization ability. These findings will provide technical support for the construction of pesticide knowledge bases. Full article

Few-shot fine-grained image classification (FSFGIC) aims to classify subspecies with similar appearances under conditions of very limited data. In this paper, we observe an interesting phenomenon: different types of image data augmentation techniques have varying effects on the performance of FSFGIC methods. This indicates that there may be biases in the features extracted from the input images. The bias of the acquired feature may cause deviation in the calculation of similarity, which is particularly detrimental to FSFGIC tasks characterized by low inter-class variation and high intra-class variation, thus affecting the classification accuracy. To address the problems mentioned, we propose an unbiased feature estimation network. The designed network has the capability to significantly optimize the quality of the obtained feature representations and effectively reduce the feature bias from input images. Furthermore, our proposed architecture can be easily integrated into any contextual training mechanism. Extensive experiments on the FSFGIC tasks demonstrate the effectiveness of the proposed algorithm, showing a notable improvement in classification accuracy. Full article

The accurate and efficient classification of steel surface defects is critical for ensuring product quality and minimizing production costs. This paper proposes a novel method based on wavelet transform and texture descriptors for the robust and precise classification of steel surface defects. By leveraging the multiscale analysis capabilities of wavelet transforms, our method extracts both broad and fine-grained textural features. It involves decomposing images using multi-level wavelet transforms, extracting a series set of statistical and textural features from the resulting coefficients, and employing Recursive Feature Elimination (RFE) to select the most discriminative features. A comprehensive series of experiments was conducted on two datasets, NEU-CLS and X-SDD, to evaluate the proposed method. The results highlight the effectiveness of the method in accurately classifying steel surface defects, outperforming the state-of-the-art techniques. Our method achieved an accuracy of 99.67% for the NEU-CLS dataset and 98.24% for the X-SDD dataset. Furthermore, we demonstrate the robustness of our method in scenarios with limited data, maintaining high accuracy, making it well-suited for practical industrial applications where obtaining large datasets can be challenging. Full article

Against the background of the sky, imaging and perception of aircraft are crucial for various vision applications. Thanks to the ever-evolving nature of the convolutional neural network (CNN), it has become easier to distinguish and recognize different types of aircraft. Nevertheless, accurate classification for sub-categories of aircraft still poses great challenges. On one hand, fine-grained recognition focuses on exploring and studying such problems. On the other hand, aircraft under different sub-categories and granularities put forward higher requirements for feature representation to classify, which led us to rethink the in-depth application of features. We noticed that information in the swin-transformer effectively represents the features in neural network layers, fully showcasing encoding and indexing for information. Through further research based on this, we proposed a better understanding of encoding and reuse for features, and innovatively performed feature coding graphically for classification. In this paper, our approach shows the effects on aircraft feature representation and classification, manifested from the flexible recognition effect at different aircraft category granularities, and outperforms other famous fine-grained classification models on this vision task. Not only did the approach we proposed demonstrate adaptability to aircraft at different classification granularities, but it also revealed the mechanisms and characteristics of feature encoding under different sample space partitions for classification. The relationship between the oriented representation of aircraft features and various classification granularities, which is manifested through different classification criteria, shows that feature coding and graph construction via the transformer opens a new door for specific defined classification tasks where objects are divided under various partition criteria, and provides another perspective on calculation and feature extraction in fine-grained classification. Full article