<![CDATA[Deep learning-based object detection for clothing and uniform items spans two research streams: fashion and apparel detection for e-commerce, and personal protective equipment (PPE) compliance detection for occupational safety. No prior review has jointly synthesized these streams with a detection-centric scope. This study addresses that gap through a PRISMA 2020-guided systematic literature review of 1,378 records from Scopus as the sole database, identified via two Boolean queries and screened for object detection or instance segmentation of clothing, uniform, or PPE items, yielding 26 empirical studies published between 2021 and 2026. One-stage YOLO-based detectors dominate the corpus (18 of 26 studies, up to 96.9% mAP50 at 7.3 ms inference), a dominance partly attributable to an architectural constraint in the PPE-domain search query; two-stage R-CNN and transformer-based architectures remain comparatively underexplored, the latter represented by only two studies. Dataset scale is markedly asymmetric between public fashion benchmarks (up to 491,895 images) and PPE-specific collections (as few as 1,330 images), though an emerging benchmark-reuse pattern partly offsets this gap within the PPE domain. This review contributes a cross-domain synthesis benchmarked against five prior reviews, an explicit architecture-dataset-metric mapping, a proposed nine-element minimum reporting standard, and a challenge taxonomy that surfaces previously undiscussed ethical and cultural-representation gaps in this literature]]>
Copyrights © 2026
