The review synthesizes 55 studies published between 2015 and 2025, spanning textile, steel, wood, ceramic, and electronics manufacturing. What sets it apart from earlier surveys is its methodology. The authors borrowed the PICOS framework—Population, Intervention, Comparator, Outcomes, Study design—from evidence-based medicine, where it is standard practice, and applied it to computer vision literature, where structured review protocols remain exceedingly rare. They extended it with a sixth dimension, Evidence Source, to assess dataset accessibility and reproducibility. Starting from 1,875 records identified in Scopus, IEEE Xplore, ScienceDirect, and SpringerLink, the team filtered down through PRISMA-guided screening to their final corpus of 55 empirical studies, each coded for material type, architecture family, baseline comparisons, performance metrics, and whether the system could actually translate detection outputs into industrial quality grades.

The architectural story the review tells is one of steady escalation. In the early period, convolutional neural networks dominated, establishing that learned representations outperform hand-crafted features in production settings. Pixel-level segmentation networks improved accuracy on microdefects, while multi-stream CNNs used parallel processing paths to capture complementary local and global features. Cascade pyramid architectures and multi-level feature fusion then made systems more resistant to variations in defect size, a persistent headache on production lines where flaws range from pinholes to sprawling dye irregularities. These convolutional foundations remain relevant today, particularly in the YOLO family of real-time detectors.

The paradigm shift arrived in 2024 and 2025 with the systematic integration of Transformer architectures. Unlike CNNs, which process local receptive fields hierarchically, Transformers employ self-attention to capture global context and long-range dependencies across entire images. The review highlights several landmark systems: a Swin Transformer-based defect detector that replaces conventional CNN backbones in Faster R-CNN, achieving 81.05 percent mean average precision on the NEU-DET steel benchmark and 99 percent on printed circuit board data; a global dual-attention Transformer that reaches an F-beta score of 0.836 on the Severstal steel dataset; and an adaptive cross-Transformer with self-supervised contrastive learning that enables few-shot adaptation to novel defect types, addressing the chronic scarcity of labeled industrial data. Meanwhile, hybrid designs flourished, pairing CNN feature extractors with Transformer attention modules to balance computational efficiency against global sensitivity.

Yet the review is refreshingly candid about the costs. Transformers carry a computational overhead of three to ten times that of optimized CNNs, a serious barrier for edge devices on resource-constrained production lines. Only two of the twelve Transformer studies the authors examined reported edge-device deployment results, raising questions about industrial readiness beyond laboratory validation. Convolutional engineering fought back impressively: YOLOv10 achieved end-to-end detection without non-maximum suppression post-processing, cutting latency by 40 percent, while YOLO-World extended detection to open vocabularies through vision-language pre-training, enabling zero-shot detection across more than 1,200 categories at 52 frames per second. The emerging consensus, the authors argue, favors hybrid architectures that synergize global modeling with convolutional inductive biases.

The review’s most original conceptual contribution is the notion of grade-readiness: the capability of a detection system to translate raw model outputs—bounding boxes, segmentation masks, confidence scores—into industrial quality classifications aligned with production standards such as ISO 12945-1 for textile pilling, ISO 21920 for metal surface texture, and ASTM specifications for stainless steel grading. The pipeline the authors envision runs from an inspection image through a detection model to multi-dimensional defect characterization—type, location, size, severity, and count—and finally to an actionable grade: Pass or Grade A, Grade B, Grade C, or Fail. The sobering finding is how rarely this bridge is actually built. Although 87.5 percent of the reviewed studies discuss industrial quality integration in their introductions, only seven of 55—about 12.5 percent—operationalize or demonstrate grade-mapping workflows. Detection metrics like mean average precision routinely exceed 0.90 in controlled benchmarks, but the translation into quality decisions that buyers and regulators recognize remains largely aspirational.

Cross-domain analysis reveals material-specific patterns that practitioners will find immediately useful. Textiles demand multi-scale architectures to accommodate diverse fabric textures and suffer from small-defect and data-imbalance problems; an enhanced YOLOv7-tiny there achieved 84 percent mAP at 90 frames per second on a proprietary dataset. Steel inspection, the dominant domain at 37.5 percent of the corpus, prioritizes real-time processing for production-line integration, with systems like IDP-Net reaching 98.7 percent mAP at 50 frames per second. Wood surfaces benefit from texture-aware features that distinguish genuine defects from natural grain variations, while particleboard grading systems have demonstrated full detection-to-grade pipelines. Ceramics and electronics remain comparatively underexplored. The catch, the authors stress, is that cross-domain comparisons are often impossible: textiles report full detection metrics while metals emphasize segmentation scores, and evaluation protocols differ so widely that state-of-the-art claims frequently cannot be tested against one another.

Reproducibility emerges as the field’s uncomfortable Achilles heel. Only about 47 percent of the datasets used across the corpus are fully public, while 14 percent are proprietary and 5 percent undisclosed. The authors’ risk-of-bias stratification sorted the 55 studies into three clusters: 78.2 percent were high quality with low bias, but 14.5 percent were high-quality studies with serious reproducibility concerns—no code repositories, no random seeds, proprietary test sets that may inflate accuracy claims—and 7.3 percent showed high risk across all six assessed bias domains and were downweighted in the synthesis. Sixty percent of studies report accuracy without any computational efficiency figures, and a quarter apply custom train-test splits to public benchmarks without disclosing split ratios or seeds. The authors recommend a minimum reporting checklist requiring both accuracy and efficiency metrics, transparent split documentation, and routine release of inference code and trained weights—practices that require no new data collection, only more complete reporting of experiments already being run.

Looking forward, the review identifies three converging directions. Self-supervised learning could reduce dependence on large annotated defect datasets by learning transferable representations from unlabeled normal-surface imagery. Standardized multimodal fusion protocols could extend the nascent RGB-depth and multi-sensor work already showing promise. And foundation models—promptable segmentation systems like SAM, vision-language models like CLIP, and industrial anomaly adaptations like AnomalyGPT—remain conspicuously absent as genuine tools in the reviewed corpus, even though they could enable zero-shot detection of defect categories no one has labeled. The authors’ verdict is that surface defect detection has matured from a research curiosity into a strategic instrument of global quality assurance, but that its promise will only be fulfilled when detection, grading, and standardized assessment are connected into a coherent, verifiable ecosystem. Fabric Quality 4.0, they conclude, must move beyond rhetoric to become an industrial reality—and this review provides both the map and the measuring stick for getting there.

Subject of Research: Automated surface defect detection and quality grading using deep learning across manufacturing industries

Article Title: Conceptualizing automated surface quality: A cross-domain evidence synthesis from CNNs to transformers

Article References: Dharma, F. P., Singgih, M. L., & Prastyo, D. D. (2026). Conceptualizing automated surface quality: A cross-domain evidence synthesis from CNNs to transformers. Machine Learning with Applications, 26, Article 100989. https://doi.org/10.1016/j.mlwa.2026.100989

Image Credits: AI Generated

DOI: 10.1016/j.mlwa.2026.100989

Keywords: surface defect detection, deep learning, convolutional neural networks, vision transformers, quality grading, textile inspection, steel inspection, systematic review, PICOS framework, YOLO, industrial automation, reproducibility

Traditional anomaly detection methodologies frequently stumble when confronted with the complexities inherent in industrial environments marked by uneven and dim illumination. Techniques relying on image enhancement often introduce artifacts or amplify noise, which can obscure critical defect features rather than clarify them. Meanwhile, deep learning models renowned for their accuracy tend to be computationally intensive and require extensive annotated datasets, limiting their practicality in fast-paced industrial scenarios. Addressing these gaps, the researchers have developed a system called DarkAD, an end-to-end framework that bypasses conventional pre-processing steps and instead integrates feature enhancement directly within the detection model.

Central to DarkAD’s innovation is the Dark-Aware Feature Adapter (DAFA), a module designed to skillfully balance noise suppression with the amplification of essential features, even under challenging lighting conditions. DAFA employs two specialized techniques: Frequency-Based Feature Enhancement (FFE) and Illumination-Aware Feature Enhancement (IFE). The FFE component adeptly distinguishes between structural information and high-frequency noise, enhancing the model’s ability to ignore spurious image artifacts. In parallel, IFE evaluates the illumination distribution across an image and dynamically strengthens the representation of poorly lit areas, ensuring that even subtle defects receive appropriate attention during analysis.

Dr. Tan elaborates on the system’s sophistication, highlighting how DarkAD circumvents the computational bottlenecks that hamper many existing approaches. “Unlike conventional models which first preprocess images with resource-heavy low-light enhancement algorithms, our framework internalizes feature extraction enhancements, enabling faster and more accurate anomaly detection. This innovation not only reduces inspection errors but also shrinks operational costs, making industrial monitoring more efficient and reliable,” he explains. This novel integration marks a significant advancement in the intersection of computer vision and industrial engineering.

The genesis of DarkAD draws from prior hybrid models such as SimpleNet, which combine feature embedding and synthesizing strategies to generate flexible anomaly detection capabilities. While SimpleNet offered computational efficiency and improved anomaly generalization, it struggled under dim lighting conditions typical of many manufacturing settings. By adapting these foundational concepts and introducing illumination-aware mechanisms, DarkAD notably surmounts limitations related to semantic inconsistencies and large data storage requirements, heralding a new era for adaptive anomaly detection.

In rigorous testing, DarkAD demonstrated remarkable resilience in pinpointing subtle defects across a diverse array of objects characterized by complex textures and shapes. Harnessing an assembled dataset painstakingly curated under low-light conditions, the researchers ensured the model learned to identify anomalies including scratches, dents, discolorations, missing components, and surface deformations prevalent in real-world industrial contexts. This meticulous approach to dataset construction amplifies the model’s real-world relevance and applicability.

The technical brilliance of the framework is further exemplified by its dynamic feature adaptation—selectively amplifying salient features from both well-lit and low-lit regions without the need for preliminary image enhancement. Such adaptability is vital in manufacturing environments where lighting can be uneven and vary dramatically across inspected surfaces. By dynamically tuning its sensitivity to illumination discrepancies, DarkAD achieves both high detection precision and speed, outperforming its predecessors.

Beyond its technical merits, the potential applications of DarkAD span a broad spectrum of industrial sectors. Real-time quality control in automotive manufacturing stands to benefit immensely, as does the production of electrical components such as cable glands and insulators. Moreover, sectors like textiles, where lighting variations and texture complexity are especially challenging, can harness this technology for improved defect detection. DarkAD’s push towards automated 24/7 monitoring also signals a future where factories and warehouses rely less on human inspectors, minimizing errors due to fatigue or environmental constraints.

The researchers emphasize that this anomaly detection approach also holds significant promise for hazardous and complex environments, such as power grid infrastructure and underwater inspection systems, where lighting is often restricted and manual inspection is impractical or dangerous. Automated monitoring powered by DarkAD promises increased safety, operational efficiency, and early anomaly detection, potentially preventing catastrophic failures and costly downtime.

Published in the journal Results in Engineering and scheduled officially for March 1, 2025, this study marks a pivotal milestone in engineering research. The work meticulously combines experimental rigor with practical engineering needs, showcasing a model that not only excels in laboratory benchmarks but is also scalable and implementable in industrial ecosystems worldwide. It represents a synthesis of computer vision, machine learning, and industrial engineering, culminating in a tool that expertly bridges the gap between theory and practice.

Dr. Tan, reflecting on the broader implications, asserts, “Our work on DarkAD underscores the vital role of customized feature enhancement for tackling the prevalent challenges posed by low-light industrial environments. This technology could revolutionize how industries perceive and respond to quality control issues, greatly reducing operational risks and propelling forward the reliability of automated systems.” Such confident projections underscore the transformative potential of this research.

In an era increasingly defined by automation and digitalization, tools like DarkAD pave the way for smarter, more reliable industrial processes. Its combination of speed, precision, and adaptability addresses a major industrial pain point while opening new frontiers in anomaly detection research. By integrating domain-specific insights with cutting-edge computer vision techniques, DarkAD exemplifies the kind of innovation capable of reshaping multiple sectors and setting new standards for quality and safety.

As industries worldwide continue to embrace smarter manufacturing and inspection methodologies, DarkAD stands out as an emblem of progress. Its real-time, adaptive detection capabilities promise not only to enhance product quality but also contribute to safer, more sustainable industrial operations in environments where light is scarce, yet vigilance remains crucial.

Subject of Research: Not applicable

Article Title: Image-based anomaly detection in low-light industrial environments with feature enhancement

News Publication Date: 1-Mar-2025

Web References: http://dx.doi.org/10.1016/j.rineng.2025.104309

References: DOI: 10.1016/j.rineng.2025.104309

Image Credits: Credit: Dr. Phan Xuan Tan from Shibaura Institute of Technology

Keywords: Applied sciences and engineering, Applied physics, Engineering, Industrial science, Information science, Risk management, Technology, Quality control, Research and development, Industrial ceramics, Metallurgy, Materials engineering, Computer processing, Automated planning, Energy infrastructure, Industrial plants, Industrial production, Systems engineering, Regulatory policy