Robust Features Research Articles

Exploring the Influence of Feature Selection Methods on a Random Forest Model for Gait Time Series Prediction using IMUs.

Despite the increasing use of IMUs and machine learning techniques for gait analysis, there remains a gap in which feature selection methods is best tailored for gait time series prediction. This study explores the impact of using various feature selection methods on the performance of a Random Forest (RF) model in predicting lower limb joints kinematics from two IMUs. This study primary objectives are: 1) Comparing eight feature selection methods based on their ability to identify more robust feature sets, time efficiency, and impact on RF models? performance, and 2) assessing the performance of RF models using generalized feature sets on a new dataset. Twenty-three typically developed children (ages 6 to 15) participated in data collection involving Optical Motion Capture, and IMUs. Joint kinematics were computed using OpenSim. By employing eight feature selection methods (four filter and four embedded methods), the study identified 30 important features for each target. These selected features were used to develop personalized and generalized RF models to predict lower limbs joints kinematics during gait. This study reveals that various feature selection methods have a minimal impact on the performance of personalized and generalized RF models. However, the RF and Mutual Information (MI) methods provided slightly lower errors and outliers. MI demonstrated remarkable robustness by consistently identifying the most common features across different participants. ElasticNet emerged as the fastest method. Overall, the study illuminated the robustness of RF models in predicting joint kinematics during gait in children, showcasing consistent performance across various feature selection methods.

Interactions between NAD+ metabolism and immune cell infiltration in ulcerative colitis: subtype identification and development of novel diagnostic models

BackgroundUlcerative colitis (UC) is a chronic inflammatory disease of the colonic mucosa with increasing incidence worldwide. Growing evidence highlights the pivotal role of nicotinamide adenine dinucleotide (NAD+) metabolism in UC pathogenesis, prompting our investigation into the subtype-specific molecular underpinnings and diagnostic potential of NAD+ metabolism-related genes (NMRGs).MethodsTranscriptome data from UC patients and healthy controls were downloaded from the GEO database, specifically GSE75214 and GSE87466. We performed unsupervised clustering based on differentially expressed NAD+ metabolism-related genes (DE-NMRGs) to classify UC cases into distinct subtypes. GSEA and GSVA identified potential biological pathways active within these subtypes, while the CIBERSORT algorithm assessed differential immune cell infiltration. Weighted gene co-expression network analysis (WGCNA) combined with differential gene expression analysis was used to pinpoint specific NMRGs in UC. Robust gene features for subtyping and diagnosis were selected using two machine learning algorithms. Nomograms were constructed and their effectiveness was evaluated using receiver operating characteristic (ROC) curves. Reverse transcription quantitative polymerase chain reaction (RT-qPCR) was conducted to verify gene expression in cell lines.ResultsIn our study, UC patients were classified into two subtypes based on DE-NMRGs expression levels, with Cluster A exhibiting enhanced self-repair capabilities during inflammatory responses and Cluster B showing greater inflammation and tissue damage. Through comprehensive bioinformatics analyses, we identified four key biomarkers (AOX1, NAMPT, NNMT, PTGS2) for UC subtyping, and two (NNMT, PARP9) for its diagnosis. These biomarkers are closely linked to various immune cells within the UC microenvironment, particularly NAMPT and PTGS2, which were strongly associated with neutrophil infiltration. Nomograms developed for subtyping and diagnosis demonstrated high predictive accuracy, achieving area under curve (AUC) values up to 0.989 and 0.997 in the training set and up to 0.998 and 0.988 in validation sets. RT-qPCR validation showed a significant upregulation of NNMT and PARP9 in inflamed versus normal colonic epithelia, underscoring their diagnostic relevance.ConclusionOur study reveals two NAD+ subtypes in UC, identifying four biomarkers for subtyping and two for diagnosis. These findings could suggest potential therapeutic targets and contribute to advancing personalized treatment strategies for UC, potentially improving patient outcomes.

A Hybrid Transfer Learning Framework for Brain Tumor Diagnosis

Brain tumors are among the most severe health challenges, necessitating early and precise diagnosis for effective treatment planning. This study introduces an optimized hybrid transfer learning (TL) framework for brain tumor classification using magnetic resonance imaging images. The proposed system integrates advanced preprocessing techniques, an ensemble of pretrained deep learning models, and explainable artificial intelligence (XAI) methods to achieve high accuracy and reliability. The methodology enhances image quality through noise reduction and contrast enhancement, facilitating robust feature extraction. The ensemble model combines VGG16 and ResNet152V2 architectures, achieving a classification accuracy of 99.47% on a challenging four‐class dataset. Additionally, gradient‐weighted class activation mapping and SHapley Additive exPlanations (SHAP)‐based XAI techniques provide visual and quantitative insights into model predictions, improving interpretability and clinical trust. This comprehensive framework demonstrates the potential of hybrid TL and XAI in advancing diagnostic accuracy and supporting clinical decision‐making for brain tumor detection. The results underscore its applicability in clinical settings, particularly in resource‐constrained environments.

The Data Artifacts Glossary: a community-based repository for bias on health datasets

BackgroundThe deployment of Artificial Intelligence (AI) in healthcare has the potential to transform patient care through improved diagnostics, personalized treatment plans, and more efficient resource management. However, the effectiveness and fairness of AI are critically dependent on the data it learns from. Biased datasets can lead to AI outputs that perpetuate disparities, particularly affecting social minorities and marginalized groups.ObjectiveThis paper introduces the “Data Artifacts Glossary”, a dynamic, open-source framework designed to systematically document and update potential biases in healthcare datasets. The aim is to provide a comprehensive tool that enhances the transparency and accuracy of AI applications in healthcare and contributes to understanding and addressing health inequities.MethodsUtilizing a methodology inspired by the Delphi method, a diverse team of experts conducted iterative rounds of discussions and literature reviews. The team synthesized insights to develop a comprehensive list of bias categories and designed the glossary’s structure. The Data Artifacts Glossary was piloted using the MIMIC-IV dataset to validate its utility and structure.ResultsThe Data Artifacts Glossary adopts a collaborative approach modeled on successful open-source projects like Linux and Python. Hosted on GitHub, it utilizes robust version control and collaborative features, allowing stakeholders from diverse backgrounds to contribute. Through a rigorous peer review process managed by community members, the glossary ensures the continual refinement and accuracy of its contents. The implementation of the Data Artifacts Glossary with the MIMIC-IV dataset illustrates its utility. It categorizes biases, and facilitates their identification and understanding.ConclusionThe Data Artifacts Glossary serves as a vital resource for enhancing the integrity of AI applications in healthcare by providing a mechanism to recognize and mitigate dataset biases before they impact AI outputs. It not only aids in avoiding bias in model development but also contributes to understanding and addressing the root causes of health disparities.

Exploring MAP3K genes in gastric cancer: biomarkers, tumor microenvironment dynamics, and chemotherapy resistance

BackgroundGastric cancer (GC) presents a significant global health burden, necessitating a deeper understanding of its molecular underpinnings for improved diagnostics and therapeutics.MethodsIn this study, we investigated the expression profiles and clinical implications of MAP3K genes in GC using in silico and in vitro experiments.ResultsUtilizing RT-qPCR analysis, we observed significant up-regulation of MAP3K1, MAP3K4, MAP3K5, MAP3K6, MAP3K7, MAP3K8, MAP3K9, and MAP3K10 in GC cell lines, while MAP3K2, MAP3K3, MAP3K11, MAP3K12, MAP3K13, MAP3K14, and MAP3K15 exhibited down-regulation. Prognostic evaluation revealed that elevated expression of MAP3K1, MAP3K4, MAP3K7, MAP3K8, MAP3K9, and MAP3K10 was associated with shorter overall survival (OS), emphasizing their clinical significance. Furthermore, the diagnostic potential was demonstrated through robust Receiver operating characteristics (ROC) curve analysis, indicating the strong discriminatory power of these genes in distinguishing GC patients. Proteomic analysis further confirmed the higher expression of MAP3K1, MAP3K4, MAP3K7, MAP3K8, MAP3K9, and MAP3K10 genes in GC. Methylation profiling further supported the idea that promoter hypomethylation of MAP3K1, MAP3K4, MAP3K7, MAP3K8, MAP3K9, and MAP3K10 genes was associated with their up-regulation. Single-cell functional analysis elucidated the involvement of MAP3K genes in shaping the tumor microenvironment. miRNA-mRNA network analysis revealed intricate regulatory mechanisms, with hsa-mir-200b-3p emerging as a key regulator. Finally, the MAP3K1 knockdown has shown significant impacts on the cellular behavior of the BGC823 cells.ConclusionThis comprehensive assessment provides valuable insights into the role of MAP3K genes in GC, offering avenues for further research and therapeutic exploration.

Microemulsion gel systems: Formulation, stability studies, biopolymer interactions, and functionality in food product development.

Microemulsion gels (MGs) are nanostructured systems created by the addition of thickening agents/biopolymers to a microemulsion's aqueous or oily phases, offering benefits like improved solubilization, enhanced stability, high encapsulation efficiency, and sustained release with versatile applications in food, pharmaceuticals, and cosmetology. MGs are intricate systems with thermodynamic robustness and controllable rheological characteristics crucial for obtaining high structural integrity and achieving innovative results regarding food product development in diverse areas of food, including colloidal carriers, food packaging, active compound delivery, antimicrobial vectors, and production of biopolymer nanoparticles. Therefore, a comprehensive analysis, hence understanding about MG systems, is needed to identify trends and gaps, helping researchers to identify promising areas for innovation and providing direction for future research. This review offers a comprehensive analysis of MG systems, their characteristics, formulation, formation mechanisms, design approaches, digestion dynamics, and rheological properties. MGs excel in solubilizing hydrophilic and lipophilic bioactives due to their enhanced viscosity and interconnected droplet network within the gel matrix. Despite their advantages, challenges, such as formulation complexity, require further understanding. This article also explores innovative biopolymers, characterization, and extensive applications, while addressing case studies, and emerging trends leveraging the potential of MG systems for enhancing food stability, functionality, and nutritional value.

Enhancing Ad Hoc Network Security using Palm Vein Biometric Features

This study proposes an innovative approach to securing ad hoc networks through palm vein biometric authentication, addressing critical security vulnerabilities in decentralized wireless communications. The research introduces an Adaptive Fusion Biometric Key Generation (AFBKG) framework that seamlessly integrates palm vein biometric features with state-of-the-art cryptographic protocols. The methodology implements a comprehensive six-stage process, incorporating Near-Infrared (NIR) imaging at 850 nm wavelength, advanced image preprocessing techniques, and deep learning-based feature extraction using a fine-tuned Convolutional Neural Network (CNN), culminating in a robust 512-dimensional feature vector. A rigorous performance evaluation was conducted, which demonstrated exceptional results, achieving 98% authentication accuracy with a 0.1% False Acceptance Rate (FAR) and 95% spoofing resistance. The AFBKG algorithm significantly outperforms traditional security methods, demonstrating 95% authentication strength and 92% resistance to Man-in-the-Middle (MITM) attacks while maintaining minimal key management complexity (15%). The system's superior scalability (90%) and computational efficiency (10% overhead) compared to conventional biometric approaches are noteworthy. These findings establish palm vein biometric authentication as a cutting-edge solution for enhancing ad hoc network security, offering substantial improvements over traditional password-based systems and alternative biometric methods.

A Novel Face Frontalization Method by Seamlessly Integrating Landmark Detection and Decision Forest into Generative Adversarial Network (GAN)

In real-world scenarios, posture variation and low-quality image resolution are two well-known factors that compromise the accuracy and reliability of face recognition system. These challenges can be overcome using various methods, including Generative Adversarial Networks (GANs). Despite this, concerns over the accuracy and reliability of GAN methods are increasing as the facial recognition market expands rapidly. The existing framework such as Two-Pathway GAN (TP-GAN) method has demonstrated that it is superior to numerous GAN methods that provide better face-texture details due to its unique deep neural network structure that allows it to perceive local details and global structure in a supervised manner. TP-GAN overcomes some of the obstacle associated with face frontalization tasks through the use of landmark detection and synthesis functions, but it remains challenging to achieve the desired performance across a wide range of datasets. To address the inherent limitations of TP-GAN, we propose a novel face frontalization method (NFF) combining landmark detection, decision forests, and data augmentation. NFF provides 2D landmark detection to integrate global structure with local details of the generator model so that more accurate facial feature representations and robust feature extractions can be achieved. NFF enhances the stability of the discriminator model over time by integrating decision forest capabilities into the TP-GAN discriminator core architecture that allows us to perform a wide range of facial pose tasks. Moreover, NFF uses data augmentation techniques to maximize training data by generating completely new synthetic data from existing data. Our evaluations are based on the Multi-PIE, FEI, and CAS-PEAL datasets. NFF results indicate that TP-GAN performance can be significantly enhanced by resolving the challenges described above, leading to high quality visualizations and rank-1 face identification.

Alpha synuclein overexpression can drive microbiome dysbiosis in mice

Growing evidence indicates that persons with Parkinson disease (PD), have a unique composition of indigenous gut microbes. Given the long prodromal or pre-diagnosed period, longitudinal studies of the human and rodent gut microbiome before symptomatic onset and for the duration of the disease are currently lacking. PD is partially characterized by the accumulation of the protein α-synuclein (α-syn) into insoluble aggregates, in both the central and enteric nervous systems. As such, several experimental rodent and non-human primate models of α-syn overexpression recapitulate some of the hallmark pathophysiologies of PD. These animal models provide an opportunity to assess how the gut microbiome changes with age under disease-relevant conditions. Here, we used a transgenic mouse strain, which overexpress wild-type human α-syn to test how the gut microbiome composition responds in this model of PD pathology during aging. Using shotgun metagenomics, we find significant, age and genotype-dependent bacterial taxa whose abundance becomes altered with age. We reveal that α-syn overexpression can drive alterations to the gut microbiome composition and suggest that it limits diversity through age. Taxa that were most affected by genotype-age interaction were Lactobacillus and Bifidobacteria. In a mouse model, we showed direct link between alpha synuclein geneotype (hallmark of PD), a dysbiotic and low-diversity gut microbiome, and dysbiotic levels of Bifidobacteria and Lactobacillus (most robust features of PD microbiome). Given emerging data on the potential contributions of the gut microbiome to PD pathologies, our data provide an experimental foundation to understand how the PD-associated microbiome may arise as a trigger or co-pathology to disease.

A robust semi-supervised regressor with correntropy-induced manifold regularization and adaptive graph.

For semi-supervised regression tasks, most existing methods ignore the impact of noise. However, the data inevitably contain noise. Therefore, this study proposes a novel correntropy-induced semi-supervised regression (CSSR) method that mitigates the adverse effects of noise. To implement the robustness of CSSR, a novel correntropy-induced manifold regularization (CMR) and a correntropy-induced adaptive graph (CAG) are designed. Specifically, CMR is inspired by the principles of correntropy and aims to learn a graph representation, whereas CAG inherits the robust characteristics of the correntropy metric and adaptively constructs an adjacency matrix. Finally, by incorporating CMR, CAG, and the correntropy-induced loss seamlessly, CSSR demonstrates the ability to deliver promising joint performance. The final solution of CSSR is achieved through an iterative process. Moreover, we validated the convergence of CSSR through a combination of theoretical analyses and empirical experiments. The experimental evaluation encompassed three synthetic, 15 benchmark, and two image datasets. The findings demonstrate that CSSR surpasses similar methods in the realm of semi-supervised regression tasks, demonstrating its effectiveness and robustness.

Applications of Algebra in Topological Data Analysis: Bridging Algebra and Data Science

Topological Data Analysis (TDA) has emerged as a powerful framework for understanding the shape and structure of data. Algebra, particularly concepts from homological and computational algebra, plays a pivotal role in TDA by enabling the extraction of robust topological features from complex datasets. This review explores the applications of algebra in TDA, highlighting its contributions to data science. We discuss foundational concepts, methodologies, computational tools, and practical applications, providing insights into the intersection of algebra and data-driven insights. We delve into the theoretical foundations of TDA, highlighting the construction of simplicial complexes, the computation of homology and persistent homology, and the development of efficient algorithms for large-scale data analysis. Furthermore, we examine the integration of TDA with machine learning, its applications across various domains including image and signal processing, natural language processing, and biosciences and discuss current challenges and future directions in the field. By bridging the gap between abstract algebraic theories and real-world data analysis, this paper underscores the transformative potential of TDA in advancing data science methodologies

MWG-UNet: Hybrid Transformer U-Net Model for Brain Tumor Segmentation in MRI Scans++

The accurate segmentation of brain tumors from medical images is critical for diagnosis and treatment planning. However, traditional segmentation methods struggle with complex tumor shapes and inconsistent image quality which leads to suboptimal results. To address this challenge, we propose multiple tasking Wasserstein Generative Adversarial Network U-shape Network++ (MWG-UNet++) to brain tumor segmentation by integrating a U-Net architecture enhanced with transformer layers which combined with Wasserstein Generative Adversarial Networks (WGAN) for data augmentation. The proposed model called Residual Attention U-shaped Network (RAUNet) for brain tumor segmentation leverages the robust feature extraction capabilities of U-Net and the global context awareness provided by transformers to improve segmentation accuracy. Incorporating WGAN for data augmentation addresses the challenge of limited medical imaging datasets to generate high-quality synthetic images that enhance model training and generalization. Our comprehensive evaluation demonstrates that this hybrid model significantly improves segmentation performance. The RAUNet outperforms compared approaches by capturing long-range dependencies and considering spatial variations. The use of WGANs augments the dataset for resulting in robust training and improved resilience to overfitting. The average evaluation metric for brain tumor segmentation is 0.8965 which outperformed the compared methods.

Unveiling encephalopathy signatures: A deep learning approach with locality-preserving features and hybrid neural network for EEG analysis.

EEG signals exhibit spatio-temporal characteristics due to the neural activity dispersion in space over the brain and the dynamic temporal patterns of electrical activity in neurons. This study tries to effectively utilize the spatio-temporal nature of EEG signals for diagnosing encephalopathy using a combination of novel locality preserving feature extraction using Local Binary Patterns (LBP) and a custom fine-tuned Long Short-Term Memory (LSTM) neural network. A carefully curated primary EEG dataset is used to assess the effectiveness of the technique for treatment of encephalopathies. EEG signals of all electrodes are mapped onto a spatial matrix from which the custom feature extraction method isolates spatial features of the signals. These spatial features are further given to the neural network, which learns to combine the spatial information with temporal dynamics summarizing pertinent details from the raw EEG data. Such a unified representation is key to perform reliable disease classification at the output layer of the neural network, leading to a robust classification system, potentially providing improved diagnosis and treatment. The proposed method shows promising potential for enhancing the automated diagnosis of encephalopathy, with a remarkable accuracy rate of 90.5%. To the best of our knowledge, this is the first attempt to compress and represent both spatial and temporal features into a single vector for encephalopathy detection, simplifying visual diagnosis and providing a robust feature for automated predictions. This advancement holds significant promise for ensuring early detection and intervention strategies in the clinical environment, which in turn enhances patient care.

Artificial Intelligence Comic Strip (AICS) Generators: A Review of Subscription Models, Pricing, and User Satisfaction

This study investigates the evolution of Artificial Intelligence Comic Strip (AICS) platforms by analyzing their pricing strategies, subscription models, and user satisfaction. Fifteen leading platforms were systematically selected based on popularity, unique features, and diverse user groups, including beginners, educators, and professionals. They employed user feedback and platform analysis to evaluate pricing models and their impact on accessibility and retention. Key findings revealed five dominant pricing strategies: freemium models, tiered subscriptions, credit-based systems, pay-per-use options, and customizable services. While freemium models effectively introduce users to the platforms, high subscription fees and complicated credit systems often hinder user retention. User feedback indicated dissatisfaction with unclear pricing structures and limited affordability, particularly in regions with financial constraints. Adaptive pricing models incorporating user engagement metrics and regional economic factors were identified as potential solutions to enhance accessibility. Platforms combining affordability with robust features achieved better satisfaction rates and broader user adoption. Additionally, the study highlighted the importance of transparent communication and simplified interfaces to address barriers for non-expert users. This study underscores the need for AICS platforms to adopt inclusive pricing and user-centric designs. By doing so, they can attract a diverse audience, encourage creative participation, and strengthen their market position. These findings contribute to understanding how AICS platforms can balance user satisfaction and revenue generation while fostering creative expression and democratizing storytelling tools.

Establishment and extension of a compact and robust binary feature descriptor for UAV image matching

Abstract Local feature description is a crucial challenge in unmanned aerial vehicle (UAV) images matching. Although scholars have explored a variety of methods for defining local feature descriptions, improving the matching accuracy of UAV images and reducing the memory consumption of descriptor remain worthwhile research issues. A compact and robust binary feature descriptor for UAV image matching is proposed. To achieve rotational invariance of the descriptor, a sampling pattern based on the region overlap error is employed to assign the dominant direction of the feature points. Additionally, a rotation invariant gradient computation method is proposed to reduce the dominant direction assignment errors. Then, the region around the feature point is divided into grids based on the region overlap error, and gradient orientation histogram and average intensity histogram are constructed. Finally, these histograms are binarized to generate a 30-byte binary descriptor. Results from the Oxford dataset, the Iguazu dataset, and a high-resolution UAV image dataset indicate that the proposed descriptor is compact and achieves the highest matching accuracy compared to nine commonly used feature descriptors, including scale-invariant feature transform, speeded up robust feature, binary robust invariant scalable keypoints, KAZE, oriented fast and rotated brief, fast retina keypoint, local intensity order pattern, boosted efficient binary local image descriptor, and triplet-based efficient binary local image descriptor. Notably, on the high-resolution UAV images with translational transformations, scale variations, salt-and-pepper noise, and Gaussian white noise, the accuracies of the proposed descriptor were 87.09%, 80.83%, 92.46%, and 87.42%, respectively. The improvements ranged from 6.95% to 53.20%, 9.32% to 55.90%, 10.39% to 44.29%, and 6.54% to 58.24%.

Robust Fine-Grained Learning for Cloth-Changing Person Re-Identification

Cloth-changing Person Re-Identification (CC-ReID) poses a significant challenge in tracking pedestrians across cameras while accounting for changes in clothing appearance. Despite recent progress in CC-ReID, existing methods predominantly focus on learning the unique biological features of pedestrians, often overlooking constraints that promote the learning of cloth-agnostic features. Addressing this limitation, we propose a Robust Fine-grained Learning Network (RFLNet) to effectively learn robust cloth-agnostic features by leveraging fine-grained semantic constraints. Specifically, we introduce a four-body-part attention module to enhance the learning of detailed pedestrian semantic features. To further strengthen the model’s robustness to clothing variations, we employ a random erasing algorithm, encouraging the network to concentrate on cloth-irrelevant attributes. Additionally, we design a fine-grained semantic loss to guide the model in learning identity-related, detailed semantic features, thereby improving its focus on cloth-agnostic regions. Comprehensive experiments on widely used CC-ReID benchmarks demonstrate the effectiveness of RFLNet. Our method achieves state-of-the-art performance, including a 0.7% increase in mAP on PRCC and a 1.6% improvement in rank-1 accuracy on DeepChange.

HFA-Net: hybrid feature-aware network for large-scale point cloud semantic segmentation

Semantic segmentation of large-scale point clouds in 3D computer vision is a challenging problem. Existing feature extraction modules often emphasize learning local geometry while not giving adequate consideration to the integration of color information. This limitation prevents the network from thoroughly learning local features, thereby impacting segmentation accuracy. In this study, we propose three modules for robust feature extraction and aggregation, forming a novel point cloud segmentation network (HFA-Net) for large-scale point cloud semantic segmentation. First, we introduce the Hybrid Feature Extraction Component (HFEC) and the Hybrid Bilateral Enhancement Component (HBAC) to comprehensively extract and enhance the geometric, color, and semantic information of point clouds. Second, we incorporate the Ternary-Distance Attention Pooling (TDAP) module, which leverages trilateral distances to further refine the network’s focus on various features, enabling it to emphasize both locally important features and broader local neighborhoods. These modules are stacked into dense residual components to expand the network’s receptive field. Our experiments on several large-scale benchmark datasets, including Semantic3D, Toronto3D, S3DIS and LASDU demonstrate the effectiveness of HFA-Net when compared to state-of-the-art networks.

A Coarse-to-Fine Feature Aggregation Neural Network with a Boundary-Aware Module for Accurate Food Recognition

Food recognition from images is crucial for dietary management, enabling applications like automated meal tracking and personalized nutrition planning. However, challenges such as background noise disrupting intra-class consistency, inter-class distinction, and domain shifts due to variations in capture angles, lighting, and image resolution persist. This study proposes a multi-stage convolutional neural network-based framework incorporating a boundary-aware module (BAM) for boundary region perception, deformable ROI pooling (DRP) for spatial feature refinement, a transformer encoder for capturing global contextual relationships, and a NetRVLAD module for robust feature aggregation. The framework achieved state-of-the-art performance on three benchmark datasets, with Top-1 accuracies of 99.80% on the Food-5k dataset, 99.17% on the Food-101 dataset, and 85.87% on the Food-2k dataset, significantly outperforming existing methods. This framework holds promise as a foundational tool for intelligent dietary management, offering robust and accurate solutions for real-world applications.

Coupling characteristics of optical skyrmions based on localized spoof plasmons (LSP)

In this work, we theoretically investigate the coupling characteristics of optical skyrmions based on localized spoof plasmons (LSPs). First, the single LSP optical skyrmion is realized and resonance modes attributed to different topological features are analyzed. Second, the coupling effect is observed through decreasing the distance between the adjacent LSP optical skyrmions. The coupling LSP optical skyrmions can preserve the topological behaviors, and particularly, three coupling modes (mode-a, mode-b, and mode-c) can be observed. Mode-a presents uniform magnetic field distributions, while mode-b and mode-c can be observed with asymmetric magnetic field distributions. Finally, the ten-coupling LSP optical skyrmions supported by space-coiling cylinders and square structures are realized and the robust topological features are discussed. The results may contribute to the future investigation on designing optical skyrmion crystals and advanced optical devices.

PBCS-ConvNeXt: Convolutional Network-Based Automatic Diagnosis of Non-alcoholic Fatty Liver in Abdominal Ultrasound Images.

Non-alcoholic fatty liver disease (NAFLD) is a highly prevalent chronic liver condition characterized by excessive hepatic fat accumulation. Early diagnosis is crucial as NAFLD can progress to more severe conditions like steatohepatitis, fibrosis, cirrhosis, and hepatocellular carcinoma without timely intervention. While liver biopsy remains the gold standard for NAFLD assessment, abdominal ultrasound (US) imaging has emerged as a widelyadopted non-invasive modality due to convenience and low cost. However, the subjective interpretation of US images is challenging and unpredictable. This study proposes a deep learning-based computer-aided diagnosis (CAD) model, termed potent boosts channel-aware separable intent - ConvNeXt (PBCS-ConvNeXt), for automated NAFLD classification using B-mode US images. The model architecture comprises three key components: The potent stemcell, an advanced trainable preprocessing module for robust feature extraction; Enhanced ConvNeXt Blocks that amplify channel-wise features to refine processing; and the boosting block that integrates multi-stage features for effective information extraction from US data. Utilizing fatty liver gradings from attenuation imaging (ATI) as the ground truth, the PBCS-ConvNeXt model was evaluated using 5-fold cross-validation, achieving an accuracy of 82%, sensitivity of 81% and specificity of 83% for identifying fatty liver on abdominal US. The proposed CAD system demonstrates high diagnostic performance in NAFLD classification from US images, enabling early detection and informing timely clinical management to prevent disease progression.