Object Detection on Crohn's Disease MRE proprietary (test)
0.5437AP @ IoU=10%DeFloMat
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DeFloMatConfidence threshold (τc)=30%, Number of proposals (Nprop)=120, Inference steps (S)=32025.12 | 0.5437 | 0.3129 | 0.4815 | 0.2937 | 0.5917 | 0.3431 | |
| DeFloMatConfidence threshold (τc)=40%, Number of proposals (Nprop)=120, Inference steps (S)=32025.12 | 0.5327 | 0.3117 | 0.5647 | 0.3424 | 0.5706 | 0.3405 | |
| DeFloMatConfidence threshold (τc)=50%, Number of proposals (Nprop)=120, Inference steps (S)=32025.12 | 0.5263 | 0.3116 | 0.5822 | 0.3553 | 0.5612 | 0.3401 | |
| DiffusionDetConfidence threshold (τc)=30%, Number of proposals (Nprop)=120, Inference steps (S)=32025.12 | 0.4913 | 0.3151 | 0.5311 | 0.3405 | 0.5175 | 0.3376 | |
| DiffusionDetConfidence threshold (τc)=40%, Number of proposals (Nprop)=120, Inference steps (S)=32025.12 | 0.4178 | 0.2779 | 0.5307 | 0.3559 | 0.4333 | 0.2924 | |
| Mask R-CNNConfidence threshold (τc)=30%2025.12 | 0.4063 | 0.2205 | 0.4041 | 0.22 | 0.4474 | 0.2478 | |
| Mask R-CNNConfidence threshold (τc)=40%2025.12 | 0.3515 | 0.2009 | 0.3889 | 0.2197 | 0.3748 | 0.2138 | |
| DiffusionDetConfidence threshold (τc)=50%, Number of proposals (Nprop)=120, Inference steps (S)=32025.12 | 0.3432 | 0.2408 | 0.4687 | 0.3295 | 0.3521 | 0.2502 | |
| Mask R-CNNConfidence threshold (τc)=50%2025.12 | 0.3092 | 0.1828 | 0.3728 | 0.2172 | 0.3239 | 0.1948 | |
| YOLOv4Confidence threshold (τc)=30%2025.12 | 0.2004 | 0.1466 | 0.2913 | 0.2126 | 0.2027 | 0.1498 | |
| YOLOv4Confidence threshold (τc)=40%2025.12 | 0.1404 | 0.1124 | 0.2185 | 0.1736 | 0.1414 | 0.1181 | |
| YOLOv4Confidence threshold (τc)=50%2025.12 | 0.0946 | 0.0769 | 0.1552 | 0.1242 | 0.0953 | 0.0777 |