Home News Latest Papers Tutorials Datasets Wiki SOTA LLM Models GPU Leaderboard Events

English

Anomaly Detection On Ubnormal

Metrics

AUC

RBDC

TBDC

Results

Performance results of various models on this benchmark

Model Name	AUC	RBDC	TBDC	Paper Title	Repository
Background-Agnostic Framework	61.3%	25.43	56.27	A Background-Agnostic Framework with Adversarial Training for Abnormal Event Detection in Video	-
TimeSformer	68.5%	0.04	0.05	Is Space-Time Attention All You Need for Video Understanding?	-
AnomalyRuler	71.9%	-	-	Follow the Rules: Reasoning for Video Anomaly Detection with Large Language Models	-
MULDE-frame-centric-micro-one-class-classification	72.8%	-	-	MULDE: Multiscale Log-Density Estimation via Denoising Score Matching for Video Anomaly Detection	-
COSKAD-euclidean	64.9%	-	-	Contracting Skeletal Kinematics for Human-Related Video Anomaly Detection	-
MIL	50.3%	0.002	0.001	Real-world Anomaly Detection in Surveillance Videos	-
STG-NF - Unsupervised	71.8%	-	-	Normalizing Flows for Human Pose Anomaly Detection	-
COSKAD-radial	62.9%	-	-	Contracting Skeletal Kinematics for Human-Related Video Anomaly Detection	-
COSKAD-hyperbolic	65%	-	-	Contracting Skeletal Kinematics for Human-Related Video Anomaly Detection	-
BiPOCO	50.7	-	-	BiPOCO: Bi-Directional Trajectory Prediction with Pose Constraints for Pedestrian Anomaly Detection	-
FPDM	62.7	-	-	Feature Prediction Diffusion Model for Video Anomaly Detection	-
SSMTL++v1	62.1%	25.63	63.53	SSMTL++: Revisiting Self-Supervised Multi-Task Learning for Video Anomaly Detection	-
STG-NF - Supervised	79.2%	-	-	Normalizing Flows for Human Pose Anomaly Detection	-
MoCoDAD	68.3%	-	-	Multimodal Motion Conditioned Diffusion Model for Skeleton-based Video Anomaly Detection	-

0 of 14 row(s) selected.

Anomaly Detection On Ubnormal | SOTA | HyperAI