Anomaly Detection on MVTec AD 8 (test) benchmark leaderboard · SOTA2 Research

Anomaly Detection on MVTec AD 8 (test)

100AUROC (Carpet)CSFlow

Evaluation Results

Method	Links
CSFlow2023.03		100	—	99	100	100	100	99.8	99.8	99.1	97.1	99.6	99.1	98.6	97.6	91.9	99.3	99.7	98.2	98.7
THFR2023.03		100	—	99.8	100	99.3	99.2	99.7	100	99.2	97.5	100	100	97.8	97.1	100	99.7	97.7	98.9	99.2
RealNet2023.03		100	—	99.9	100	100	99.2	99.8	100	99.2	99.5	100	99.8	99.1	98.8	99.4	100	99.8	99.6	99.6
HETMM2023.03		100	—	100	100	100	99.7	99.9	100	100	100	100	100	98.2	99.2	100	100	99.8	99.7	99.8
UniAD2023.03		99.9	—	99.8	100	99	97.9	99.1	100	97.6	85.3	99.9	99	88.3	91.9	95	100	96.7	95.4	96.6
EffiAD2023.03		99.9	—	100	100	100	99.5	99.7	100	95.2	97.9	99.4	99.6	98.6	97	100	99.9	99.7	98.7	99.1
MSFlow2023.03		99.8	—	97	100	99.9	99.8	99.8	100	99.3	99.3	100	100	98.7	97.6	100	100	99.8	99.5	99.6
HETMM2023.03		99.8	—	99.9	100	99.9	99.7	99.9	100	100	100	100	100	97.5	95.8	100	100	99.2	99.2	99.5
HVQT2023.03		99.3	—	98.4	100	99.2	97.2	98.7	100	99	95.4	100	99.9	95.8	95.6	93.6	99.7	97.9	97.7	98
RD4AD2023.03		98.9	—	100	100	99.3	99.2	99.5	100	95	96.3	99.9	100	96.6	97	99.5	96.7	98.5	98	98.5
PNI2023.03		98.5	—	100	100	100	99.6	99.6	100	99.8	99.7	100	100	96.7	99.5	99.7	100	99.9	99.5	99.5
PatchCore2023.03		98	—	95.6	100	99.4	99.2	99.1	100	99.3	98	100	99.7	97	96.4	100	99.9	99.2	99	99.1