Search Results (2,474)

Accurate cultivar identification of commercial dried goji berries is essential for raw material sorting, batch consistency assessment, and quality control during processing and distribution. Conventional approaches based on manual judgment or physicochemical analysis are often subjective, labor-intensive, time-consuming, and costly, making them unsuitable for rapid commercial sorting and quality inspection. To develop a rapid, low-cost, and nondestructive method for dried goji berry cultivar identification, this study proposes a visual recognition framework that integrates RGB imaging with lightweight deep learning. A dataset comprising 25,899 RGB images from five cultivars of commercial dried goji berry samples, namely Ningqi No. 7, Linqi No. 5, Ningqi No. 1, Keqi 6082, and Jingqi No. 1, was constructed. Given the pronounced surface shrinkage, complex texture, and subtle inter-cultivar appearance differences of dried goji berries, an image quality enhancement method was designed to strengthen the representation of color gradation, textural details, and edge information. For model development, CoAtNet was selected as the baseline network and redesigned for lightweight deployment. By integrating an improved feature extraction module and an information-preserving downsampling module, the proposed LSH-CoAtNet model enhances fine-grained feature representation while reducing computational cost. On the quality-enhanced image dataset, the proposed method achieved an accuracy of 98.80%, a precision of 98.81%, a recall of 98.80%, and an F1-score of 98.80%. The model contained only 6.41 M parameters and required 1.60 GFLOPs, outperforming the baseline model in both classification performance and computational efficiency. Ablation experiments and five-fold cross-validation further confirmed the effectiveness of the image quality enhancement strategy, the contribution of each improved module, and the stability of the model. Overall, the proposed method, which combines RGB image quality enhancement with LSH-CoAtNet, provides a low-cost, nondestructive, and efficient technical solution for rapid cultivar identification, raw material sorting, batch consistency assessment, and quality control of commercial dried goji berries during processing and distribution. It may also serve as a reference for intelligent classification and quality inspection of other specialty dried horticultural products. Full article

This paper proposes the Continuous Operational Feedback Loop (COFL) architecture, a fully localized, event-driven operational monitoring and response system for Direct-to-Consumer (D2C) commerce. The architecture integrates the n8n workflow engine with on-premise large language model (LLM) inference via the Ollama framework, forming a containerized stack deployable on commodity CPU-only edge hardware (~USD 1640). Using a multi-source dataset of 1800 records constructed from publicly available e-commerce corpora and evaluated with a silver-standard automated labeling protocol, empirical validation demonstrates an end-to-end latency of 3.22 s and a macro-F1 sentiment classification score of 0.836—representing 98.2% of the full-precision baseline and 94.0% of cloud GPT-4o API generation quality measured by ROUGE-L—at approximately 1/200th of the per-request inference cost. A systematic quantization ablation study across six model-quantization configurations establishes LLaMA 3 8B Q4_K_M as the Pareto-optimal selection for the target hardware. An Analytic Hierarchy Process (AHP) multi-criteria framework with criterion weights derived from published literature confirms the COFL implementation achieves a higher composite score than cloud API deployment under the stated evaluation assumptions. Failure mode and effects analysis (FMEA) is summarized to characterize system reliability under identified failure scenarios. Full article

Sleep-related respiratory disturbances are difficult to monitor continuously outside specialized laboratories because conventional polysomnography is resource-intensive and intrusive. This study presents a contactless edge-AI engineering prototype for detecting controlled voluntary respiratory-motion suppression and motion artifacts using a 60 GHz frequency-modulated continuous-wave radar. The system integrates a 60 GHz radar front end, lightweight local preprocessing, an INT8 one-dimensional convolutional neural network deployed on the Analog Devices MAX78000 CNN accelerator (Analog Devices Thailand, Chon Buri, Thailand), and an event-driven Raspberry Pi Zero 2W gateway for alert transmission. Evaluation was performed using a controlled healthy-volunteer dataset consisting of normal breathing, voluntary breath-holding-induced respiratory suppression, and deliberate motion artifact. The final valid test set contained 270 technically valid 30 s windows balanced across the three classes. The INT8 model achieved an overall accuracy of 92.6% (95% confidence interval: 88.8–95.2%), with a macro-averaged precision, recall, and F1-score of 92.6%, 92.6%, and 92.5%, respectively. Active CNN inference on the MAX78000 consumed 0.152 ± 0.011 mJ and was completed in 5.20 ± 0.11 ms, corresponding to approximately 280-fold lower active inference energy than Python 3.14.6/TensorFlow Lite 2.21.0-based execution on the Raspberry Pi Zero 2W. These results demonstrate the feasibility of privacy-aware, low-power respiratory-pattern classification at the edge. However, the study should be interpreted strictly as an engineering proof-of-concept based on controlled voluntary breathing and movement tasks in healthy volunteers. It is not a clinically validated apnea or obstructive sleep apnea detection system and did not include polysomnography, oxygen saturation measurement, airflow sensing, sleep staging, or diagnosed patient cohorts. Full article

The digitization of existing building stock often depends on legacy 2D raster floor plans (scanned drawings, PDF exports, or photographs) because structured building information models are frequently unavailable for older properties. Manual measurement and visual inspection of such documents are time consuming and error prone. This paper presents an integrated deep learning pipeline that extracts semantic information from unstructured two-dimensional floor plan images of existing structures and supports preliminary compliance screening via locally deployed large language models. The pipeline employs YOLOv8 for the localization and classification of 18 architectural symbols and furniture items, and a U-Net with a ResNet34 encoder for the semantic segmentation of walls and interior room spaces. To translate pixel-level predictions into physical metrics, we implement an area calculation module based on user-defined reference scale calibration. An LLM evaluation module, deployed locally via Ollama with a retrieval-augmented generation pipeline, interprets extracted room metrics and flags potential non-compliance against referenced residential design guidelines; it is intended for the assessment of existing layouts rather than generative co-design. We expand a core dataset of 101 manually annotated source floor plans to 303 augmented instances using label-aligned geometric transformations, while reporting generalization in terms of the 101 unique source plans. On the held-out validation split (10 source plans), YOLOv8 achieves 92.3% mAP50 versus 87.2% for a Faster R-CNN reference model on the same data split (detection baselines differ in training epochs and pretraining; see Experiments); U-Net achieves 95.71% mIoU, surpassing DeepLabv3+ (93.2%) under matched segmentation training settings. The system is deployed as an interactive web application for legacy building survey and preliminary regulatory review when only two-dimensional documentation is available. Full article

Intracranial hemorrhage (ICH) is a life-threatening neurological condition that requires rapid and accurate diagnosis to reduce mortality and improve patient health. Computed tomography (CT) imaging is widely used for ICH detection. However, manual interpretation can be time-consuming and prone to errors, particularly in high-volume clinical settings. Recent studies have demonstrated the effectiveness of deep learning techniques in automating medical image analysis and improving diagnostic accuracy. In this study, we propose a novel deep learning model, MGiT-X, for the automated detection of intracranial hemorrhage using head CT images. The MGiT-X model is a hybrid deep learning architecture that uses dual scale Swin Transformer modules to extract features at multiple scales, capturing local and global contextual information on CT images. It has a Gradient Fusion mechanism to enhance feature representation by combining complementary features to distinguish between hemorrhagic and healthy tissue. In addition, to further improve feature representation, the use of Jensen–Shannon divergence is used to provide better mutual alignment and consistency between the distribution of features. An adaptive weight strategy is also employed to provide refinement to the importance of features for classification. MGiT-X is evaluated on two publicly available datasets including the Head CT Hemorrhage dataset and the Brain CT Hemorrhage dataset. The proposed approach leverages advanced feature extraction and classification capabilities to distinguish between hemorrhage and healthy cases effectively. Experimental results demonstrate that the proposed MGiT-X achieves high performance across both datasets. On Dataset 1, the model attains an overall accuracy of 95.87% and a Kappa score of 91.80%, while on Dataset 2, it achieves an improved accuracy of 99.12% with a Kappa score of 98.20%. Class-wise evaluation further shows strong performance, with F1-scores exceeding 95% for both hemorrhage and healthy classes across datasets. Full article

Background: Respiratory diseases (RDs), including asthma, COVID-19, chronic obstructive pulmonary disease (COPD), and pneumonia, remain a major global health challenge, contributing substantially to global morbidity and mortality. Conventional diagnosis relies heavily on clinicians’ expertise to interpret respiratory sounds and radiographic images, a process that can be subjective, time-consuming, and prone to inter-observer variability. Recent advances in artificial intelligence (AI) and machine learning (ML) have enabled automated diagnostic approaches that can improve the efficiency, consistency, and scalability of respiratory disease detection. However, existing research remains fragmented across different data modalities. Methods: This review systematically analyzes recent studies on AI-based respiratory disease detection using both visual modalities (e.g., chest X-rays, computed tomography (CT) scans, and ultrasound) and audio modalities (e.g., cough and breath sounds). To provide a comprehensive perspective, the reviewed literature is organized using a unified taxonomy that categorizes existing approaches into three main groups: audio-based, visual-based, and audio–visual-based methods. In addition, two conceptual frameworks are proposed to illustrate representative pipelines for audio-based and visual-based respiratory disease classification. Results: The analysis reveals that most existing studies focus on single-modality approaches, while multimodal integration remains relatively underexplored. Only a limited number of studies combine audio and visual data within unified frameworks, primarily due to the scarcity of synchronized multimodal datasets collected from the same patients. The proposed taxonomy and conceptual frameworks provide a structured basis for comparing existing methods, identifying methodological trends, and highlighting key research gaps in multimodal respiratory disease detection. Conclusions: Future research should prioritize the development of multimodal datasets, robust evaluation protocols, and interpretable and lightweight AI models suitable for real-world clinical deployment. Advancing multimodal integration has the potential to significantly enhance the accuracy, reliability, and clinical applicability of AI-driven respiratory disease diagnosis systems. Full article

Extended producer responsibility (EPR) is intended to link producer design decisions to end-of-life costs, but collective EPR schemes typically weaken this link by routing funding through producer responsibility organisations. We develop a multi-level framework of consumer-integrated environmental protection (CIEP) and argue that individual producer responsibility (IPR), where producers bear product-specific end-of-life liability, can function as a governance mechanism that reconnects design, consumer behaviour and waste governance. This paper is a qualitative multiple-case research study—not a systematic review—which draws on three funded research projects: (i) small and medium-sized enterprise (SME) tools for design-for-recyclability, (ii) an artificial intelligence (AI) application for household waste sorting, and (iii) closed-loop recycling of fishing gear in Vietnam. Within the first project (ToCoReRaM), a PRISMA-based systematic review of web-accessible circular economy tools finds that only 2 of 23 tools are SME-accessible through standard web searches. The AI-based waste-sorting application achieves approximately 75% classification accuracy under real-world conditions. The fishing gear study demonstrates technical and economic viability of closed-loop recycling, and a survey of more than 1500 Vietnamese fishers finds 95.8% willingness to return used gear given appropriate incentives. Together, the cases show that effective circular governance requires four complementary elements: IPR-based producer accountability, SME-accessible design tools, digital consumer guidance at the point of disposal, and context-sensitive governance capacity. These findings inform policy pathways for Sustainable Development Goal (SDG) 12 and SDG 14. Full article

Manual compliance auditing in cloud environments consumes up to 40% of IT security budgets annually, yet existing approaches verify control presence rather than effectiveness, leaving institutions vulnerable to adversarial evasion. This paper presents an AI-augmented hybrid ML–LLM compliance auditing system evaluated on Rwanda’s National Cyber Security Authority (NCSA) Minimum Cybersecurity Standards (169 controls across 14 families). The system combines leakage-free XGBoost multi-label classification with GPT-4o-mini semantic log analysis, grounded in a formal effectiveness model. Key findings: (1) XGBoost v2 achieves 85.45% macro-F1 on leakage-free synthetic data (Wilson 95% CI = [84.9%, 86.0%]); an initial 86.3% data-leakage rate artificially inflated prior results to 99.99% and was identified and corrected in this revision; (2) GPT-4o-mini achieves 92.3% macro accuracy across four log types (n = 628, 37.5% real enterprise data, Wilson CI = [89.9%, 94.3%]); (3) adversarial validation across five MITRE ATT&CK scenarios yields 92.8% macro detection with 0.0% false-positive rate on real SSH/PAM compliant logs (n = 75); (4) a cross-dataset generalization analysis confirms 87.6% F1 on real SSH logs but identifies a 37.8-percentage-point out-of-vocabulary gap for Windows and HTTP log types, motivating the hybrid architecture; (5) the combined hybrid system (XGBoost for in-vocabulary logs, GPT-4o-mini for out-of-vocabulary) achieves 85.1% F1 with 6.4% false-positive rate on 180 real-world logs. The system runs at 2.0 CPU cores, 2.66 GB RAM, on $50/month cloud hosting (Apple M1 Pro baseline; storage and maintenance excluded), producing audit reports in 2–5 s depending on log volume and policy document size, demonstrating that effectiveness-based compliance auditing is accessible without enterprise-grade infrastructure. Full article

Accurate traffic sign detection is important for the safety of autonomous driving systems. However, fully supervised methods require a large amount of manual annotation, which is cost-prohibitive and time-consuming. Semi-supervised methods employ a small amount of labeled data and a large amount of unlabeled data to train the models, hence largely reducing the annotation costs. However, these methods have the following challenges: (1) with an imbalanced long-tail class distribution of traffic signs, they tend to achieve poor performance on tail classes; (2) they often fail to detect small traffic signs. To solve these issues, we propose a Semi-Supervised Traffic Sign Detection method with Dynamic Pseudo-Label Selection and Gated Feature Fusion-based Proposal Refinement. Firstly, we design a Class Distribution-based Dynamic Pseudo-Label Selection module (CD-DPLS) to select pseudo-labels for different classes based on the class distribution information, which reduces the tendency to select more pseudo-labels from head classes instead of tail classes, thereby improving the tail class detection performance. Secondly, we employ a Gated Feature Fusion-based Proposal Refinement strategy (GFF-PR) to refine detection proposals by fusing different-scale features with a gating mechanism, which facilitates the detection of small traffic signs. In addition, we use an Adaptive-Weight Focal Loss (AWFL), with which the weight of each pseudo-label is determined by the ratio between its classification confidence and the corresponding class-specific classification-confidence threshold. Experiments on traffic sign datasets demonstrate that the proposed method outperforms state-of-the-art semi-supervised approaches, with mAP₅₀ scores of 10.8% and 34.9% using only 1% and 10% labeled data, respectively. Full article

Background: Allergic contact dermatitis (ACD) is a common inflammatory skin disease and patch testing (PT) remains the gold standard for its diagnosis; however, PT interpretation is time-consuming and prone to inter-observer variability. Growing advances in digital imaging and artificial intelligence (AI) have encouraged the development of automated PT evaluation systems. This review aimed to summarize the use of deep learning networks (DNNs) and image-preprocessing techniques for PT classification. Methods: A literature review was conducted to identify original research published between 2020 and 2025 that applied deep learning algorithms to PT image analysis. Included studies were assessed with respect to model architecture, dataset characteristics, preprocessing strategies, and diagnostic performance. Results: Six original studies employing deep learning for PT image classification met the inclusion criteria. They employed a range of architectures, including YOLOv5x, EfficientNetB0, Xception, and custom CNN models. Reported diagnostic performance varied, with accuracy values ranging from 90% to 99.5%, F1-scores from 0.37 to 0.98, and AUROC values up to 0.94. Despite promising results, models remain unreliable for ICDRG grading, especially for severe reactions, and methodological variability in dataset composition, imaging conditions, preprocessing pipelines, and classification tasks limits comparability across studies. Conclusions: Deep learning shows promise for automated PT interpretation, but further standardized and multicenter studies with detailed preprocessing protocols and comprehensive ICDRG grading are required for clinical implementation. Full article

Machine learning (ML) has substantially advanced text classification by enabling automated understanding and categorization of complex unstructured textual data. However, accurately capturing nuanced linguistic patterns and contextual variations that are inherent in natural language, particularly within consumer complaints, remains a challenge. This study addresses these limitations by incorporating human-experience-trained algorithms that effectively recognize subtle semantic distinctions that are crucial for assessing consumer relief eligibility. Specifically, we propose synthetic data generation methods that leverage generative adversarial networks refined through expert labeling, complemented by featurization approaches for extracting textual representations. By combining expert-trained classifiers with high-quality synthetic data, this research demonstrates marked improvements in ML classifier performance, reduced dataset acquisition costs, and enhanced robustness across evaluation metrics in text classification tasks. Full article

Image sensors produce high-dimensional visual data for classification algorithms. Deep Neural Networks (DNNs) achieve high accuracy but require large labeled datasets and computational and energy resources, limiting their use in embedded systems. Active Learning (ALrn) can reduce labeling effort by selecting samples based on informativeness scores, but it remains computationally expensive, especially for high-dimensional images. This work presents a hardware-accelerated approach for the instance selection stage based on a query strategy in uncertainty-based ALrn for image classification using a novel in-line top-k selection algorithm that avoids conventional sorting and reduces memory and computational requirements. The algorithm is implemented on an Xilinx ZYNQ-7000 System on Chip (SoC) using a Field Programmable Gate Array (FPGA)-based accelerator operating at 110 MHz, interfacing with an embedded Advanced RISC Machine (ARM) processor for data acquisition and communication via the Python Productivity for Zynq (PYNQ) framework. Experiments on diverse multiclass datasets demonstrate correctness within an ALrn setting, showing negligible performance deviation in the learning curves compared to software baselines. The accelerator achieves speedup of 231.7× and 22.9× over software baseline and optimized software implementation of the proposed algorithm, respectively, in query-strategy computation while consuming only 0.473 W, substantially lower than conventional Central Processing Unit (CPU)- and Graphics Processing Unit (GPU)-based platforms. These results demonstrate the efficiency and extensibility of the proposed accelerator across alternative ALrn designs and hardware platforms, where the computational cost of instance selection scales with the size of the unlabeled pool. Full article

The integration of renewable energy and rising electricity demand strain system flexibility. While price-based demand response (PBDR) improves flexibility through pricing signals, its efficacy hinges critically on accurate consumer modeling. Recognizing this pivotal role, this paper provides a comprehensive review of consumer models in PBDR and their applications to electric vehicles (EVs). First, a unified conceptual framework is presented, delineating the energy, information and financial flows among the system operator (SO), load aggregators (LAs), and end-users, and highlighting the central position of consumer modeling. Second, existing modeling approaches are systematically classified into four categories, namely rule-based, optimization-based, data-driven, and hybrid, to facilitate the selection of appropriate models by researchers and stakeholders for diverse scenarios. Furthermore, the application and adaptation of these models to EVs are critically analyzed, accounting for unique vehicular constraints. Subsequently, a systematic summary of the key characteristics and existing research gaps is provided. Finally, key directions for future research are proposed accordingly, aimed at incorporating bounded rationality into behavioral models, developing individualized consumer modeling coupled with user-specific dynamic pricing, and extending consumer modeling to residential multi-energy prosumers in integrated energy systems. Full article

From the perspective of applied surface science, heritage textiles from the Kashgar region can be regarded as fragile fibrous surface systems in which stains, abrasion, yarn breakage, yarn shedding, holes, and color fading represent measurable surface-degradation phenomena. However, manual inspection of these complex, woven, embroidered, and aged surfaces is time-consuming and difficult to standardize. To support non-contact surface-condition documentation, this study proposes an improved YOLOv8-based framework, YOLOv8-MABFT, for surface defect detection and condition-level assessment of Kashgar heritage textiles. The model integrates the C2f-Faster-EMA module and an RT-DETR-informed decoder head to improve the detection of weak-boundary and fine-grained surface defects. A dataset of 8247 high-resolution annotated images was constructed, covering six surface-degradation categories: stains, broken yarn, yarn shedding, holes, abrasion, and color fading. Experimental results show that YOLOv8-MABFT achieves an F1-score of 94.6%, a precision of 91.4%, a recall of 98.0%, and an mAP@0.5 of 94.0%, outperforming Faster R-CNN, SSD, YOLOv5n, YOLOv7n, and YOLOv8n while maintaining lightweight computational characteristics. CAM-based visualizations indicate that the improved model focuses more consistently on defect-related surface regions rather than surrounding decorative textures. Based on detected defects, seven surface-condition variables were constructed and input into a Random Forest classifier for four-level condition prediction. SHAP analysis shows that Distribution and Severity are the main contributors to condition classification. Overall, the proposed framework provides an applied surface-science tool for non-contact surface defect detection, surface-condition documentation, and preliminary condition-level assessment of fragile textile substrates. Full article

The rapid growth of user-generated reviews on e-commerce platforms has created a significant decision-making challenge for both consumers and sellers, particularly in morphologically rich low-resource languages such as Turkish. This study proposes a unified artificial intelligence framework for Turkish e-commerce review intelligence by integrating transformer-based sentiment classification, instruction-tuned large language model summarization, and explainable fuzzy logic-based product evaluation within a single end-to-end architecture. A balanced dataset containing 183,333 Turkish reviews was constructed from Trendyol, Amazon Turkey, and Hepsiburada using LLM-assisted annotation and stratified downsampling. Experimental evaluations demonstrated that the fine-tuned BERTurk 128k model achieved a macro F1-score of 0.9243 on the held-out test set. To overcome the limitations of multilingual news-oriented summarization models on informal review text, the framework employed the Turkish instruction-tuned Kumru-2B model together with structured prompt engineering to generate sentiment-aware abstractive summaries. In addition, a Mamdani-type fuzzy inference system was designed to combine sentiment distribution, seller reliability, star ratings, and review volume into an interpretable product-level score. The complete pipeline was integrated into a FastAPI and React-based web platform capable of processing approximately 850 reviews in under 60 s. The findings demonstrate that domain-specific Turkish language models combined with explainable reasoning mechanisms can provide accurate, scalable, and human-interpretable decision support for large-scale e-commerce environments. Full article