Anomaly Detection 벤치마크
Anomaly Detection on Lost and Found
AP
- 2021-03-09 — SynBoost: AP 70.43
- 2021-11-24 — PEBAL: AP 78.29
- 2023-07-25 — Mask2Anomaly: AP 86.59
| Rank | Model | AP | FPR | Extra Training Data | Paper | Code | Year |
|---|---|---|---|---|---|---|---|
| 1 | Mask2Anomaly | 86.59 | 5.75 | ✓ | Unmasking Anomalies in Road-Scene Segmentation | shyam671/mask2anomaly-unmasking-anomalies-in-road-scene-segmentation | 2023 |
| 2 | PEBAL | 78.29 | 0.81 | ✓ | Pixel-wise Energy-biased Abstention Learning for Anomaly Segmentation on Complex Urban Driving Scenes | tianyu0207/pebal · yyliu01/rpl · gaozhitong/atta | 2021 |
| 3 | SynBoost | 70.43 | 4.89 | ✓ | Pixel-wise Anomaly Detection in Complex Driving Scenes | giandbt/synboost | 2021 |
| 4 | SML | 25.89 | 44.48 | Standardized Max Logits: A Simple yet Effective Approach for Identifying Unexpected Road Obstacles in Urban-Scene Segmentation | shjung13/standardized-max-logits | 2021 |