Detect image anomalies from many classes using one model with UniAD
Detect image anomalies from many classes using one model with UniAD
A Unified Model for Multi-class Anomaly Detection
arXiv paper abstract https://arxiv.org/abs/2206.03687v1
arXiv PDF paper https://arxiv.org/pdf/2206.03687v1.pdf
Despite the rapid advance of unsupervised anomaly detection, existing methods require to train separate models for different objects.
… present UniAD that accomplishes anomaly detection for multiple classes with a unified framework.
… reconstruction networks may fall into an “identical shortcut”, where both normal and anomalous samples can be well recovered, and hence fail to spot outliers.
To tackle … First … come up with a layer-wise query decoder to help model the multi-class distribution. Second, … employ a neighbor masked attention module …
Third, … propose a feature jittering strategy that urges the model to recover the correct message even with noisy inputs.
… algorithm … surpass the state-of-the-art alternatives by a sufficiently large margin …
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
