Anomaly Detection On Shanghaitech
평가 지표
AUC
평가 결과
이 벤치마크에서 각 모델의 성능 결과
비교 표
| 모델 이름 | AUC |
|---|---|
| follow-the-rules-reasoning-for-video-anomaly | 85.2% |
| videopatchcore-an-effective-method-to | 85.1% |
| an-exploratory-study-on-human-centric-video | 80.6% |
| bounding-boxes-and-probabilistic-graphical | 61.28% |
| eval-explainable-video-anomaly-localization | 76.63% |
| ubnormal-new-benchmark-for-supervised-open | 83.7% |
| learning-regularity-in-skeleton-trajectories | 73.40% |
| video-anomaly-detection-by-solving-decoupled | 84.3% |
| multi-timescale-trajectory-prediction-for | 76.03% |
| any-shot-sequential-anomaly-detection-in | 71.6% |
| self-supervised-predictive-convolutional | 83.6% |
| ssmtl-revisiting-self-supervised-multi-task | 83.8% |
| stan-spatio-temporal-adversarial-networks-for | 76.2% |
| attention-based-residual-autoencoder-for | 73.6 |
| context-recovery-and-knowledge-retrieval-a | 83.7% |
| a-revisit-of-sparse-coding-based-anomaly | 68.0% |
| spatio-temporal-predictive-tasks-for-abnormal | 77.1% |
| diversity-measurable-anomaly-detection | 78.8% |
| divide-and-conquer-in-video-anomaly-detection | 87.72% |
| mulde-multiscale-log-density-estimation-via | 86.7% |
| attribute-based-representations-for-accurate | 85.94% |
| mulde-multiscale-log-density-estimation-via | 81.3% |
| ssmtl-revisiting-self-supervised-multi-task | 82.9% |
| anomaly-detection-in-video-via-self | 82.4% |
| regularity-learning-via-explicit-distribution | 83.35 |
| a-scene-agnostic-framework-with-adversarial | 82.7% |
| self-supervised-masked-convolutional | 83.6% |
| self-supervised-predictive-convolutional | - |
| making-anomalies-more-anomalous-video-anomaly | 76.5% |
| object-centric-auto-encoders-and-dummy | 78.7% |
| normalizing-flows-for-human-pose-anomaly | 85.9% |