Anomaly Localization on MVTec-AD 2019 benchmark leaderboard · SOTA2 Research

Anomaly Localization on MVTec-AD 2019

99BottleP-Net (P-NET*)

Evaluation Results

Method
P-Net (P-NET*)2020.07	99	70	84	57	98	97	89	79	91	100	97	99	82	98	90	89	0.12
FCDD2020.07	97	90	93	96	91	95	98	94	81	86	91	94	88	88	92	92	0.04
FCDD2020.07	96	93	95	99	95	97	99	98	97	93	98	95	90	94	98	96	0.02
Self-Similarity (AE-SS*)2020.07	93	82	94	87	94	97	78	89	91	96	59	92	90	73	88	86	0.1
Gradient Descent Reconstruction with VAEs (GDR*)2020.07	92	91	92	74	96	98	93	91	93	95	65	99	92	84	87	89	0.09
Visually Explained Variational Autoencoder (VEVAE*)2020.07	87	90	74	78	73	98	95	94	83	97	80	94	93	77	78	86	0.09
L2 Autoencoder (AE-L2*)2020.07	86	86	88	59	90	95	75	86	85	96	51	93	86	73	77	82	0.13
AnoGAN*2020.07	86	78	84	54	58	87	64	76	87	80	50	90	80	62	78	74	0.13
Superpixel Masking and Inpainting (SMAI*)2020.07	86	92	93	88	97	97	86	92	92	96	62	96	85	80	90	89	0.09
CNN Feature Dictionaries (CNNFD*)2020.07	78	79	84	72	59	72	87	82	68	87	93	77	66	91	76	78	0.1