Reproduce “Anomaly Clustering: Grouping Images into Coherent Clusters of Anomaly Types” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Kihyuk Sohn, Jinsung Yoon, Chun-Liang Li, Chen-Yu Lee, Tomas Pfister
The authors present an anomaly clustering framework that groups product images using weighted local patch features. It represents each image as a bag of pretrained features and clusters images by distances between weighted summaries. On MVTec object categories, it reports 0.451 NMI, versus 0.244 for uniform averaging; results are from a transductive benchmark with a known cluster count.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.