HiProtoAD: Feature Reconstruction Guided by Cluster-Separated Hierarchical Prototypes for Multi-Class Unsupervised Anomaly Detection
IEEE Access · 2026 · 14 · 99798-99815
DOI: 10.1109/access.2026.3707880 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Chaoming Li und 2 weiteren Autor:innen. Die Arbeit nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Young Teachers Scientific Research Funding Project of Zhaoqing University.
Abstract (Originalsprache)
Prevailing Transformer-based feature reconstruction methods for multi-class unsupervised visual anomaly detection suffer from identity mapping, while the underlying causes remain insufficiently understood. Through a systematic analysis, we reveal that identity mapping is closely related to anomaly injection arising from cross-category information interaction and model architectures. To address this issue, we propose HiProtoAD, a novel framework that guides feature reconstruction via cluster-separated hierarchical prototype learning. Built upon a Vision Transformer–based hierarchical reconstruction architecture, HiProtoAD incorporates three key components: 1) a Cluster-Separated Learnable Token Selector (CSLTS) to isolate category-specific normal prototype extraction and mitigate cross-category anomaly injection; 2) a Sparse Cross-Attention (SCA) module to suppress low-relevance anomalous information; and 3) a Gaussian-Masked Cross-Attention (GMCA) module to enhance normal prototype-guided feature reconstruction. Extensive experiments on benchmark datasets, including MVTec-AD, VisA, BTAD, and MPDD, demonstrate that HiProtoAD achieves competitive performance compared with state-of-the-art methods, providing an effective solution for complex multi-class unsupervised visual anomaly detection.
Förderung
- Young Teachers Scientific Research Funding Project of Zhaoqing University
Zitieren
APA
Li, C., Yu, G., & Ye, W. (2026). HiProtoAD: Feature Reconstruction Guided by Cluster-Separated Hierarchical Prototypes for Multi-Class Unsupervised Anomaly Detection. IEEE Access, 14, 99798-99815. https://doi.org/10.1109/access.2026.3707880
Vancouver
Li C, Yu G, Ye W. HiProtoAD: Feature Reconstruction Guided by Cluster-Separated Hierarchical Prototypes for Multi-Class Unsupervised Anomaly Detection. IEEE Access. 2026;14:99798-99815. doi:10.1109/access.2026.3707880
BibTeX
@article{Li2026,
author = {Li, Chaoming and Yu, Guoqing and Ye, Wenwen},
title = {HiProtoAD: Feature Reconstruction Guided by Cluster-Separated Hierarchical Prototypes for Multi-Class Unsupervised Anomaly Detection},
journal = {IEEE Access},
year = {2026},
volume = {14},
pages = {99798--99815},
doi = {10.1109/access.2026.3707880}
}
EUVIMED