AI Paper Atlas

Research map / Label-efficient and robust learning / OOD and anomaly detection

Divide-and-Assemble: Learning Block-Wise Memory for Unsupervised Anomaly Detection

Jinlei Hou, Yingying Zhang, Qiaoyong Zhong, Di Xie, Shiliang Pu, Hong Zhou

ICCV 2021 · Label-efficient and robust learning › OOD and anomaly detection · approach: Unified multi-class anomaly detection · 215 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add pixel-wise memory module (MemAE) (ablation)AE · MVTec ADAUROC: 0.794 -> 0.802No clear effect
add plain skip connections (ablation)AE · MVTec ADAUROC: 0.794 -> 0.736Hurt
use multi-scale block-wise memory module (ablation)AE · MVTec ADAUROC: 0.794 -> 0.845Improved
add adversarially learned representation (ablation)DAAD · MVTec ADAUROC: 0.845 -> 0.895Improved
set division rate to 8 rather than 1 (ablation)block-wise memory · MVTec ADAUROC: 0.716 -> 0.845Improved
increase division rate from 8 to 16 (ablation)block-wise memory · MVTec ADAUROC: 0.845 -> 0.813Hurt
add adversarial loss alone (ablation)DAAD · MVTec ADAUROC: 0.845 -> 0.847No clear effect
add alignment loss (ablation)DAAD · MVTec ADAUROC: 0.847 -> 0.855Improved
add alignment score L(x) to anomaly score (ablation)DAAD · MVTec ADAUROC: 0.855 -> 0.874Improved
fuse reconstruction score R(x) with L(x) (ablation)DAAD · MVTec ADAUROC: 0.874 -> 0.895Improved
add adversarial representation components (ablation)AE · MVTec ADAUROC: 0.794 -> 0.835Improved
add adversarial representation components (ablation)MemAE · MVTec ADAUROC: 0.802 -> 0.853Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy