Object Detection on People-Art (test)
68AP50StyleCOCO
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| StyleCOCONumber of training images (n)=58,6722021.02 | 68 | 0.36 | 0.33 | |
| Fast R-CNNBackbone=VGG16, Training=fine-tuned on People-Art2021.02 | 58 | — | — | |
| Multiple Instance Learning (MIL)Task=Weakly Supervised Object Detection (WSOD)2021.02 | 58 | — | — | |
| YOLOTraining=not fine-tuned on People-Art2021.02 | 45 | — | — | |
| Deformable Parts Model (DPM)2021.02 | 40 | — | — | |
| Fast R-CNNNet=VGG16, Training Mode=Fine tuned2018.10 | — | 59 | — | |
| MAXNet=RES-152-COCO2018.10 | — | 25.9 | — | |
| MI-maxNet=RES-152-COCO2018.10 | — | 55.4 | — | |
| MI-SVMNet=RES-152-COCO2018.10 | — | 13.3 | — |