Benchmark Datasets Research Articles

Promoting camouflaged object detection through novel edge-target interaction and frequency-spatial fusion

The realm of camouflage object detection endeavors to precisely discern concealed targets that blend seamlessly into their surroundings. Recently, numerous research endeavors have converged on unraveling the disguise tactics employed by such objects, striving for efficient and accurate detection results. While these methods have garnered notable advancements in unveiling camouflaged targets, they still grapple with challenges posed by the high similarity between objects and their backgrounds, resulting in misidentifications, detection oversights, and the loss of intricate details. In light of these challenges, this paper introduces an innovative EFNet, aimed at elevating the integrity of target recognition and the finesse of detail capture through deepened interactions between Edge-target features and Frequency-spatial information. Notably, recognizing that the edges of camouflaged objects harbor a wealth of intrinsic details and serve as crucial delimiters distinguishing them from their surroundings, we have devised a dual-path architecture, where an auxiliary Edge Detection (ED) branch collaborates seamlessly with the primary Object Segmentation (OS) branch. Within the OS branch, we meticulously crafted the Edge-induced Segmentation Refinement module (ESR), which ingeniously harnesses the refined edge information provided by the ED branch to bolster target segmentation accuracy and augment detail fidelity. Moreover, in the supporting ED branch, we propose the Structure-induced Edge Enhancement module (SEE), designed to accentuate edge information while suppressing irrelevant noise with the assistance of the OS branch, thereby ensuring the precise extraction of edge information. To further enrich the information on dual-path, we innovatively introduce a Frequency Decomposition Module (FDM), which decomposes images into low- and high-frequency components, tailored to enhance target segmentation and edge detection, respectively. Additionally, to ensure seamless integration of frequency-domain and spatial-domain information, we devised the versatile Frequency-Spatial Domain Mixer (FSDM), achieving precise information alignment and fusion of these two domains. Quantitative and qualitative experimental results demonstrate that our proposed EFNet significantly outperforms existing state-of-the-art methods on four widely used benchmark datasets.

An Exploratory Review on Recent Computational Approaches Devised for MiRNA Disease Association Prediction

Recent evidence demonstrated the fundamental role of miRNAs as disease biomarkers and their role in disease progression and pathology. Identifying disease related miRNAs using computational approaches has become one of the trending topics in health informatics. Many biological databases and online tools were developed for uncovering novel disease-related miRNAs. Hence, a brief overview regarding the disease biomarkers, miRNAs as disease biomarkers and their role in complex disorders is given here. Various methods for calculating miRNA and disease similarities are included and the existing machine learning and network based computational approaches for detecting disease associated miRNAs are reviewed along with the benchmark dataset used. Finally, the performance matrices, validation measures and online tools used for miRNA Disease Association (MDA) predictions are also outlined.

Information gap based knowledge distillation for occluded facial expression recognition

Facial Expression Recognition (FER) with occlusion presents a challenging task in computer vision because facial occlusions result in poor visual data features. Recently, the region attention technique has been introduced to address this problem by researchers, which make the model perceive occluded regions of the face and prioritize the most discriminative non-occluded regions. However, in real-world scenarios, facial images are influenced by various factors, including hair, masks and sunglasses, making it difficult to extract high-quality features from these occluded facial images. This inevitably limits the effectiveness of attention mechanisms. In this paper, we observe a correlation in facial emotion features from the same image, both with and without occlusion. This correlation contributes to addressing the issue of facial occlusions. To this end, we propose a Information Gap based Knowledge Distillation (IGKD) to explore the latent relationship. Specifically, our approach involves feeding non-occluded and masked images into separate teacher and student networks. Due to the incomplete emotion information in the masked images, there exists an information gap between the teacher and student networks. During training, we aim to minimize this gap to enable the student network to learn this relationship. To enhance the teacher’s guidance, we introduce a joint learning strategy where the teacher conducts knowledge distillation on the student during the training of the teacher. Additionally, we introduce two novel constraints, called knowledge learn and knowledge feedback loss, to supervise and optimize both the teacher and student networks. The reported experimental results show that IGKD outperforms other algorithms on four benchmark datasets. Specifically, our IGKD achieves 87.57% on Occlusion-RAF-DB, 87.33% on Occlusion-FERPlus, 64.86% on Occlusion-AffectNet, and 73.25% on FED-RO, clearly demonstrating its effectiveness and robustness.

ATP_mCNN: Predicting ATP binding sites through pretrained language models and multi-window neural networks.

Adenosine triphosphate plays a vital role in providing energy and enabling key cellular processes through interactions with binding proteins. The increasing amount of protein sequence data necessitates computational methods for identifying binding sites. However, experimental identification of adenosine triphosphate-binding residues remains challenging. To address the challenge, we developed a multi-window convolutional neural network architecture taking pre-trained protein language model embeddings as input features. In particular, multiple parallel convolutional layers scan for motifs localized to different window sizes. Max pooling extracts salient features concatenated across windows into a final multi-scale representation for residue-level classification. On benchmark datasets, our model achieves an area under the ROC curve of 0.95, significantly improving on prior sequence-based models and outperforming convolutional neural network baselines. This demonstrates the utility of pre-trained language models and multi-window convolutional neural networks for advanced sequence-based prediction of adenosine triphosphate-binding residues. Our approach provides a promising new direction for elucidating binding mechanisms and interactions from primary structure.

Loss function inversion for improved crack segmentation in steel bridges using a CNN framework

Automating bridge visual inspection using deep learning algorithms for crack detection in images is a prominent way to make these inspections more effective. This paper addresses several challenges associated with crack detection: (1) data imbalance, caused by a small crack area as compared to the background, and (2) a high false positive rate, due to a large amount of crack-like features in the background. First, a new benchmark dataset is presented, containing images of cracks in steel bridges along with pixel-wise annotations. Secondly, the importance of incorporating background patches is examined to assess their impact on network performance when applied to high resolution images of cracks in steel bridges. Finally, a loss function is introduced that enables the use of a relatively large number of background patches in neural network training. The proposed approaches yield a significant reduction in false positive rates, thereby improving the overall performance of crack segmentation.

STCNet: Spatio-Temporal Cross Network with subject-aware contrastive learning for hand gesture recognition in surface EMG.

This paper introduces the Spatio-Temporal Cross Network (STCNet), a novel deep learning architecture tailored for robust hand gesture recognition in surface electromyography (sEMG) across multiple subjects. We address the challenges associated with the inter-subject variability and environmental factors such as electrode shift and muscle fatigue, which traditionally undermine the robustness of gesture recognition systems. STCNet integrates a convolutional-recurrent architecture with a spatio-temporal block that extracts features over segmented time intervals, enhancing both spatial and temporal analysis. Additionally, a rolling convolution technique designed to reflect the circular band structure of the sEMG measurement device is incorporated, thus capturing the inherent spatial relationships more effectively. We further propose a subject-aware contrastive learning framework that utilizes both subject and gesture label information to align the representation of vector space. Our comprehensive experimental evaluations demonstrate the superiority of STCNet under aggregated conditions, achieving state-of-the-art performance on benchmark datasets and effectively managing the variability among different subjects. The implemented code can be found at https://github.com/KNU-BrainAI/STCNet.

Signed graph embedding via multi-order neighborhood feature fusion and contrastive learning.

Signed graphs have been widely applied to model real-world complex networks with positive and negative links, and signed graph embedding has become a popular topic in the field of signed graph analysis. Although various signed graph embedding methods have been proposed, most of them still suffer from the generality problem. Namely, they cannot simultaneously achieve the satisfactory performance in multiple downstream tasks. In view of this, in this paper we propose a signed embedding method named MOSGCN which exhibits two significant characteristics. Firstly, MOSGCN designs a multi-order neighborhood feature fusion strategy based on the structural balance theory, enabling it to adaptively capture local and global structure features for more informative node representations. Secondly, MOSGCN is trained by using the signed graph contrastive learning framework, which further helps it learn more discriminative and robust node representations, leading to the better generality. We select link sign prediction and community detection as the downstream tasks, and conduct extensive experiments to test the effectiveness of MOSGCN on four benchmark datasets. The results illustrate the good generality of MOSGCN and the superiority by comparing to state-of-the-art methods.

XMEN: a modular toolkit for cross-lingual medical entity normalization.

To improve performance of medical entity normalization across many languages, especially when fewer language resources are available compared to English. We propose xMEN, a modular system for cross-lingual (x) medical entity normalization (MEN), accommodating both low- and high-resource scenarios. To account for the scarcity of aliases for many target languages and terminologies, we leverage multilingual aliases via cross-lingual candidate generation. For candidate ranking, we incorporate a trainable cross-encoder (CE) model if annotations for the target task are available. To balance the output of general-purpose candidate generators with subsequent trainable re-rankers, we introduce a novel rank regularization term in the loss function for training CEs. For re-ranking without gold-standard annotations, we introduce multiple new weakly labeled datasets using machine translation and projection of annotations from a high-resource language. xMEN improves the state-of-the-art performance across various benchmark datasets for several European languages. Weakly supervised CEs are effective when no training data is available for the target task. We perform an analysis of normalization errors, revealing that complex entities are still challenging to normalize. New modules and benchmark datasets can be easily integrated in the future. xMEN exhibits strong performance for medical entity normalization in many languages, even when no labeled data and few terminology aliases for the target language are available. To enable reproducible benchmarks in the future, we make the system available as an open-source Python toolkit. The pre-trained models and source code are available online: https://github.com/hpi-dhc/xmen.

GPLM: Enhancing underwater images with Global Pyramid Linear Modulation

Underwater imagery often suffers from challenges such as color distortion, low contrast, blurring, and noise due to the absorption and scattering of light in water. These degradations complicate visual interpretation and hinder subsequent image processing. Existing methods struggle to effectively address the complex, spatially varying degradations without prior environmental knowledge or may produce unnatural enhancements.To overcome these limitations, we propose a novel method called Global Pyramid Linear Modulation that integrates physical degradation modeling with deep learning for underwater image enhancement. Our approach extends Feature-wise Linear Modulation to a four-dimensional structure, enabling fine-grained, spatially adaptive modulation of feature maps. Our method captures multi-scale contextual information by incorporating a feature pyramid architecture with self-attention and feature fusion mechanisms, effectively modeling complex degradations. We validate our method by integrating it into the MixDehazeNet model and conducting experiments on benchmark datasets. Our approach significantly improves the Peak Signal-to-Noise Ratio, increasing from 28.6 dB to 30.6 dB on the EUVP-515-test dataset. Compared to recent state-of-the-art methods, our method consistently outperforms them by over 3 dB in PSNR on datasets with ground truth. It improves the Underwater Image Quality Measure by more than one on datasets without ground truth. Furthermore, we demonstrate the practical applicability of our method on a real-world underwater dataset, achieving substantial improvements in image quality metrics and visually compelling results. These experiments confirm that our method effectively addresses the limitations of existing techniques by adaptively modeling complex underwater degradations, highlighting its potential for underwater image enhancement tasks.

LCDL: Classification of ICD codes based on disease label co-occurrence dependency and LongFormer with medical knowledge.

Medical coding involves assigning codes to clinical free-text documents, specifically medical records that average over 3,000 markers, in order to track patient diagnoses and treatments. This is typically accomplished through manual assignments by healthcare professionals. To improve efficiency and accuracy while reducing the workload on these professionals, researchers have employed a multi-label classification approach. Since the long-tail phenomenon impacts tens of thousands of ICD codes, whereby only a few codes (representative of common diseases) are frequently assigned, while the majority of codes (representative of rare diseases) are infrequently assigned, this paper presents an LCDL model that addresses the challenge at hand by examining the LongFormer pre-trained language model and the disease label co-occurrence map. To enhance the performance of automated medical coding in the biomedical domain, hierarchies with medical knowledge, synonyms and abbreviations are introduced, improving the medical knowledge representation. Test evaluations are extensively conducted on the benchmark dataset MIMIC-III, and obtained the competitive performance compared to the previous state-of-the-art methods.

EPFDNet: Camouflage object detection with edge perception in frequency domain

Camouflaged object detection (COD) is a relatively new field of computer vision research. The challenge of this task lies in accurately segmenting camouflaged objects from backgrounds that are similar in appearance. In fact, the generation of reliable edges is an effective mean of distinguishing between the foreground and background of the image, which is beneficial for assisting in determining the location of camouflaged objects. Inspired by this, we design an Edge Encoder that decomposes features into different frequency bands adopting learnable wavelets and focuses on high-frequency components with sufficient edge details to extract accurate edges. Subsequently, the Feature Aggregation Module is proposed to integrate contextual features, which generates rough edge details by sensing the difference between two branch features and use this information to further refine our edge features. Furthermore, the Stage Enhancement Module is developed to enhance the features through reverse attention guidance and dilate convolution, which mines the detailed structural information of the camouflaged objects area by eliminating foreground. The superiority of our proposed method (EPFDNet) over the existing 17 state-of-the-art methods is demonstrated through extensive experiments on three widely used COD benchmark datasets. The code has been released at https://github.com/LitterMa-820/EPFDNet.

PyDeid: an improved, fast, flexible, and generalizable rule-based approach for deidentification of free-text medical records.

Deidentification of personally identifiable information in free-text clinical data is fundamental to making these data broadly available for research. However, there exist gaps in the deidentification landscape with regard to the functionality and flexibility of extant tools, as well as suboptimal tradeoffs between deidentification accuracy and speed. To address these gaps and tradeoffs, we develop a new Python-based deidentification software, pyDeid. pyDeid uses a combination of regular expression-based rules, fixed exclusion lists and inclusion lists to deidentify free-text data. Additional configurations of pyDeid include optional named entity recognition and custom name lists. We measure its deidentification performance and speed on 700 admission notes from a Canadian hospital, the publicly available n2c2 benchmark dataset of American discharge notes, as well as a synthetic dataset of artificial intelligence (AI) generated admission notes. We also compare its performance with the Physionet De-identification Software and the popular open-source Philter tool. Different configurations of pyDeid outperformed other tools on various metrics, with a "best" accuracy value of 0.988, best precision of 0.889, best recall of 0.950, and best F1 score of 0.904. All configurations of pyDeid were significantly faster than Philter and Physionet De-identification Software, with the fastest deidentification speed of 0.48s per note. pyDeid allows the flexibility to prioritize between performance and speed, as well as precision and recall, while addressing some of the gaps in functionality left by other tools. pyDeid is also generalizable to domains outside of clinical data and can be further customized for specific contexts or for particular workflows.

Select & Re-Rank: Effectively and efficiently matching multimodal data with dynamically evolving attention

Extracting semantically consistent representations from multi-modal data helps computers understand the human world more comprehensively. Visual-semantic matching, as one of the fundamental tasks for multi-modal learning, attracts continuous attention. Recent research makes unflagging endeavors to enhance the matching performance, but sometimes at the expense of overlooking the delicate balance between efficiency and effectiveness. In this paper, we aim to address this dilemma through a newly proposed attention-mechanism-based architecture. To ensure optimal effectiveness, we adopt a more advanced Transformer Encoder (TE) as our basic model and introduce two significant ameliorations to tailor it for the visual-semantic matching task. Initially, we incorporate fine-grained supervision into the classic TE, allowing our model to capture sophisticated correspondences between different modalities. Subsequently, we employ a dynamic attention-evolving strategy to selectively pass useful information and strengthen the attention pattern consistency between adjacent TE blocks. To maintain efficiency, we propose a novel Select & Re-rank strategy that enables the model to ignore redundant information. This approach significantly reduces the computational cost and increases the matching speed with relatively minimal performance degradation. The proposed architecture can gradually capture and reorganize useful information from inter-modality as well as intra-modality under the supervision of both fine-grained and global similarity, which leads to more comprehensive and discriminative embeddings. Experiments on two benchmark datasets show that the proposed method achieves competitive results in terms of both efficiency and effectiveness.

MutualDTA: An Interpretable Drug-Target Affinity Prediction Model Leveraging Pretrained Models and Mutual Attention.

Efficient and accurate drug-target affinity (DTA) prediction can significantly accelerate the drug development process. Recently, deep learning models have been widely applied to DTA prediction and have achieved notable success. However, existing methods often encounter several common issues: first, the data representations lack sufficient information; second, the extracted features are not comprehensive; and third, most methods lack interpretability when modeling drug-target binding. To overcome the above-mentioned problems, we propose an interpretable deep learning model called MutualDTA for predicting DTA. MutualDTA leverages the power of pretrained models to obtain accurate representations of drugs and targets. It also employs well-designed modules to extract hidden features from these representations. Furthermore, the interpretability of MutualDTA is realized by the Mutual-Attention module, which (i) establishes relationships between drugs and proteins from the perspective of intermolecular interactions between drug atoms and protein amino acid residues and (ii) allows MutualDTA to capture the binding sites based on attention scores. The test results on two benchmark data sets show that MutualDTA achieves the best performance compared to the 12 state-of-the-art models. Attention visualization experiments show that MutualDTA can capture partial interaction sites, which not only helps drug developers reduce the search space for binding sites, but also demonstrates the interpretability of MutualDTA. Finally, the trained MutualDTA is applied to screen high-affinity drug screens targeting Alzheimer's disease (AD)-related proteins, and the screened drugs are partially present in the anti-AD drug library. These results demonstrate the reliability of MutualDTA in drug development.

Classification of Imagined Speech Signals Using Functional Connectivity Graphs and Machine Learning Models.

EEG involves recording electrical activity generated by the brain through electrodes placed on the scalp. Imagined speech classification has emerged as an essential area of research in brain-computer interfaces (BCIs). Despite significant advances, accurately classifying imagined speech signals remains challenging due to their complex and non-stationary nature. Existing approaches often struggle with low signal-to-noise ratios and high inter-subject variability. A proposed method named imagined speech functional connectivity graph (ISFCG) is implemented to deal with these issues. The functional connectivity graphs capture the complex relationships between brain regions during imagined speech tasks. These graphs are then used to extract features that serve as inputs to various machine-learning models. The ISFCG provides an alternative representation of imagined speech signals, focusing on brain connectivity features to enhance the analysis and classification process. Also, a convolutional neural network (CNN) is proposed to learn features from these complex graphs, leading to improved classification accuracy. Experimental results on a benchmark dataset demonstrate the effectiveness of our method.

Image Segmentation Framework for Detecting Adversarial Attacks for Autonomous Driving Cars

The widespread deployment of deep neural networks (DNNs) in critical real-time applications has spurred significant research into their security and robustness. A key vulnerability identified is that DNN decisions can be maliciously altered by introducing carefully crafted noise into the input data, leading to erroneous predictions. This is known as an adversarial attack. In this paper, we propose a novel detection framework leveraging segmentation masks and image segmentation techniques to identify adversarial attacks on DNNs, particularly in the context of autonomous driving systems. Our defense technique considers two levels of adversarial detection. The first level mainly detects adversarial inputs with large perturbations using the U-net model and one-class support vector machine (SVM). The second level of defense proposes a dynamic segmentation algorithm based on the k-means algorithm and a verifier model that controls the final prediction of the input image. To evaluate our approach, we comprehensively compare our method to the state-of-the-art feature squeeze method under a white-box attack, using eleven distinct adversarial attacks across three benchmark and heterogeneous data sets. The experimental results demonstrate the efficacy of our framework, achieving overall detection rates exceeding 96% across all adversarial techniques and data sets studied. It is worth mentioning that our method enhances the detection rates of FGSM and BIM attacks, reaching average detection rates of 95.65% as opposed to 62.63% in feature squeezing across the three data sets.

Wavelet-Based, Blur-Aware Decoupled Network for Video Deblurring

Video deblurring faces a fundamental challenge, as blur degradation comprehensively affects frames by not only causing detail loss but also severely distorting structural information. This dual degradation across low- and high-frequency domains makes it challenging for existing methods to simultaneously restore both structural and detailed information through a unified approach. To address this issue, we propose a wavelet-based, blur-aware decoupled network (WBDNet) that innovatively decouples structure reconstruction from detail enhancement. Our method decomposes features into multiple frequency bands and employs specialized restoration strategies for different frequency domains. In the low-frequency domain, we construct a multi-scale feature pyramid with optical flow alignment. This enables accurate structure reconstruction through bottom-up progressive feature fusion. For high-frequency components, we combine deformable convolution with a blur-aware attention mechanism. This allows us to precisely extract and merge sharp details from multiple frames. Extensive experiments on benchmark datasets demonstrate the superior performance of our method, particularly in preserving structural integrity and detail fidelity.

Towards a Unified Temporal and Event Logic Paradigm for Multi-Hop Path Reasoning in Knowledge Graphs

Path reasoning in knowledge graphs is a pivotal task for uncovering complex relational patterns and facilitating advanced inference processes. It also holds significant potential in domains such as power electronics, where real-time reasoning over dynamic, evolving data is essential for advancing topology design and application systems. Despite its importance, traditional approaches often encounter substantial limitations when applied to dynamic, time-sensitive scenarios. These models typically fail to adequately capture intricate logical dependencies and demonstrate suboptimal performance in data-constrained environments. To address these challenges, we introduce Path-Reasoning Logic (PRlogic), an innovative framework that seamlessly integrates rule-based logical reasoning with cutting-edge neural network methodologies. PRlogic enhances path inference by leveraging a context-aware logical association network adept at handling temporal and event-driven attributes, enabling improved reasoning for dynamic systems such as IoT-based power electronics and smart grids. This adaptability allows the framework to better accommodate evolving knowledge structures, significantly improving reasoning accuracy under resource-scarce conditions. Furthermore, PRlogic employs a multi-stage refinement strategy, harmonizing logic-based rules with learned contextual representations to achieve heightened robustness and scalability. Comprehensive experiments on widely-recognized benchmark datasets validate the superiority of PRlogic, demonstrating its consistent outperformance of existing models in path reasoning tasks. These results underscore the efficacy of incorporating logic-driven mechanisms into knowledge graph reasoning and highlight PRlogic’s potential as a powerful solution for applications in dynamic data environments.

Semantic-Guided Transformer Network for Crop Classification in Hyperspectral Images

The hyperspectral remote sensing images of agricultural crops contain rich spectral information, which can provide important details about crop growth status, diseases, and pests. However, existing crop classification methods face several key limitations when processing hyperspectral remote sensing images, primarily in the following aspects. First, the complex background in the images. Various elements in the background may have similar spectral characteristics to the crops, and this spectral similarity makes the classification model susceptible to background interference, thus reducing classification accuracy. Second, the differences in crop scales increase the difficulty of feature extraction. In different image regions, the scale of crops can vary significantly, and traditional classification methods often struggle to effectively capture this information. Additionally, due to the limitations of spectral information, especially under multi-scale variation backgrounds, the extraction of crop information becomes even more challenging, leading to instability in the classification results. To address these issues, a semantic-guided transformer network (SGTN) is proposed, which aims to effectively overcome the limitations of these deep learning methods and improve crop classification accuracy and robustness. First, a multi-scale spatial–spectral information extraction (MSIE) module is designed that effectively handle the variations of crops at different scales in the image, thereby extracting richer and more accurate features, and reducing the impact of scale changes. Second, a semantic-guided attention (SGA) module is proposed, which enhances the model’s sensitivity to crop semantic information, further reducing background interference and improving the accuracy of crop area recognition. By combining the MSIE and SGA modules, the SGTN can focus on the semantic features of crops at multiple scales, thus generating more accurate classification results. Finally, a two-stage feature extraction structure is employed to further optimize the extraction of crop semantic features and enhance classification accuracy. The results show that on the Indian Pines, Pavia University, and Salinas benchmark datasets, the overall accuracies of the proposed model are 98.24%, 98.34%, and 97.89%, respectively. Compared with other methods, the model achieves better classification accuracy and generalization performance. In the future, the SGTN is expected to be applied to more agricultural remote sensing tasks, such as crop disease detection and yield prediction, providing more reliable technical support for precision agriculture and agricultural monitoring.

Enhancing Zero-Shot Learning Through Kernelized Visual Prototypes and Similarity Learning

Zero-shot learning (ZSL) holds significant promise for scaling image classification to previously unseen classes by leveraging previously acquired knowledge. However, conventional ZSL methods face challenges such as domain-shift and hubness problems. To address these issues, we propose a novel kernelized similarity learning approach that reduces intraclass similarity while increasing interclass similarity. Specifically, we utilize kernelized ridge regression to learn visual prototypes for unseen classes in the semantic vectors. Furthermore, we introduce kernel polarization and autoencoder structures into the similarity function to enhance discriminative ability and mitigate the hubness and domain-shift problems. Extensive experiments on five benchmark datasets demonstrate that our method outperforms state-of-the-art ZSL and generalized zero-shot learning (GZSL) methods, highlighting its effectiveness in improving classification performance for unseen classes.