Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (BTF)

Reported on 3 benchmarks across 1 task · 1 paper · 3 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology3 results

Anomaly DetectiononMVTEC 3D-AD
Detection AUCROC· uses extra data· 2022-03-10
0.865
best: 0.982 (TransFusion)
SOTA
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection arXiv:2203.05550
Anomaly DetectiononMVTEC 3D-AD
Segmentation AUCROC· uses extra data· 2022-03-10
0.992
best: 0.996 (Shape-Guided)
SOTA
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection arXiv:2203.05550
Anomaly DetectiononMVTEC 3D-AD
Segmentation AUPRO· uses extra data· 2022-03-10
0.959
best: 93.75 (CDO)
SOTA
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection arXiv:2203.05550