Anomaly Detection on MVTEC 3D-AD

Metric: Segmentation AUROC (higher is better)

LeaderboardDataset

Loading chart...

Results

Hide augmentations

Sort:

#	Model↕	Segmentation AUROC▼	Augmentations	Paper	Date↕	Code
1	AST	97.6	Yes	Asymmetric Student-Teacher Networks for Industri...	2022-10-14	Code
2	GLFM	0.979	No	-	-	Code
3	CPMF (2D+3D)	0.9781	Yes	Complementary Pseudo Multimodal Feature for Poin...	2023-03-23	Code
4	CPMF (2D+3D)	0.9781	Yes	Complementary Pseudo Multimodal Feature for Poin...	2023-03-23	Code
5	Shape-Guided (only SDF)	0.978	No	-	-	Code
6	CPMF (3D)	0.978	Yes	Complementary Pseudo Multimodal Feature for Poin...	2023-03-23	Code
7	Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (FPFH)	0.978	No	Back to the Feature: Classical 3D Features are (...	2022-03-10	Code
8	CPMF (3D)	0.978	Yes	Complementary Pseudo Multimodal Feature for Poin...	2023-03-23	Code
9	Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (SIFT)	0.974	No	Back to the Feature: Classical 3D Features are (...	2022-03-10	Code
10	CPMF (2D)	0.973	Yes	Complementary Pseudo Multimodal Feature for Poin...	2023-03-23	Code
11	CPMF (2D)	0.973	Yes	Complementary Pseudo Multimodal Feature for Poin...	2023-03-23	Code
12	MC4AD	0.946	Yes	Examining the Source of Defects from a Mechanica...	2025-05-09	Code
13	Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (HoG)	0.93	No	Back to the Feature: Classical 3D Features are (...	2022-03-10	Code
14	Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (Depth iNet)	0.93	Yes	Back to the Feature: Classical 3D Features are (...	2022-03-10	Code
15	Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (NSA)	0.817	No	Back to the Feature: Classical 3D Features are (...	2022-03-10	Code
16	Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (RaW)	0.771	No	Back to the Feature: Classical 3D Features are (...	2022-03-10	Code

#1ASTSOTA
97.6
Segmentation AUROC· Augmentations· 2022-10-14
Asymmetric Student-Teacher Networks for Industrial Anomaly Detection Code
#2GLFM
0.979
Segmentation AUROC
No paperCode
#3CPMF (2D+3D)
0.9781
Segmentation AUROC· Augmentations· 2023-03-23
Complementary Pseudo Multimodal Feature for Point Cloud Anomaly Detection Code
#4CPMF (2D+3D)
0.9781
Segmentation AUROC· Augmentations· 2023-03-23
Complementary Pseudo Multimodal Feature for Point Cloud Anomaly Detection Code
#5Shape-Guided (only SDF)
0.978
Segmentation AUROC
No paperCode
#6CPMF (3D)
0.978
Segmentation AUROC· Augmentations· 2023-03-23
Complementary Pseudo Multimodal Feature for Point Cloud Anomaly Detection Code
#7Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (FPFH)SOTA
0.978
Segmentation AUROC· 2022-03-10
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection Code
#8CPMF (3D)
0.978
Segmentation AUROC· Augmentations· 2023-03-23
Complementary Pseudo Multimodal Feature for Point Cloud Anomaly Detection Code
#9Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (SIFT)
0.974
Segmentation AUROC· 2022-03-10
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection Code
#10CPMF (2D)
0.973
Segmentation AUROC· Augmentations· 2023-03-23
Complementary Pseudo Multimodal Feature for Point Cloud Anomaly Detection Code
#11CPMF (2D)
0.973
Segmentation AUROC· Augmentations· 2023-03-23
Complementary Pseudo Multimodal Feature for Point Cloud Anomaly Detection Code
#12MC4AD
0.946
Segmentation AUROC· Augmentations· 2025-05-09
Examining the Source of Defects from a Mechanical Perspective for 3D Anomaly Detection Code
#13Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (HoG)
0.93
Segmentation AUROC· 2022-03-10
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection Code
#14Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (Depth iNet)
0.93
Segmentation AUROC· Augmentations· 2022-03-10
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection Code
#15Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (NSA)
0.817
Segmentation AUROC· 2022-03-10
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection Code
#16Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection (RaW)
0.771
Segmentation AUROC· 2022-03-10
Back to the Feature: Classical 3D Features are (Almost) All You Need for 3D Anomaly Detection Code