Research Article
Object Detection Based on Swin Deformable Transformer-BiPAFPN-YOLOX
Table 5
Experimental results of various scales networks on the COCO 2017 test set. Swin DeTr denotes Swin Deformable Transformer.
| | Models | AP (%) | AP50 | AP75 | APS | APM | APL | Param (M) | GFLOPs | Infer time (ms) | FPS |
| | DarkNet53-PAFPN-YOLOX-S [2] | 39.6 | 64.6 | 47.5 | 22.7 | 48.4 | 54.1 | 9.0 | 26.8 | 9.8 | 102.0 | | Swin DeTr-BiPAFPN-YOLOX-S | 44.7 | 67.7 | 50.3 | 25.9 | 50.9 | 59.6 | 7.1 | 16.3 | 8.2 | 122.4 | | DarkNet53-PAFPN-YOLOX-M [2] | 46.4 | 65.4 | 50.6 | 26.3 | 51.0 | 59.9 | 25.3 | 73.8 | 12.3 | 81.3 | | Swin DeTr-BiPAFPN-YOLOX-M | 48.4 | 69.3 | 53.7 | 28.7 | 52.4 | 61.2 | 21.2 | 50.6 | 9.6 | 104.3 | | DarkNet53-PAFPN-YOLOX-L [2] | 50.0 | 68.5 | 54.5 | 29.8 | 54.5 | 64.4 | 68.2 | 195.6 | 14.5 | 69.0 | | Swin DeTr-BiPAFPN-YOLOX-L | 51.8 | 69.6 | 55.4 | 31.7 | 55.8 | 66.0 | 63.5 | 181.7 | 10.7 | 93.7 | | DarkNet53-PAFPN-YOLOX-X [2] | 51.2 | 69.6 | 55.7 | 31.2 | 56.1 | 66.1 | 99.1 | 286.9 | 17.3 | 57.8 | | Swin DeTr-BiPAFPN-YOLOX-X | 52.1 | 70.4 | 57.8 | 31.9 | 56.9 | 66.7 | 85.5 | 225.4 | 14.1 | 71.0 |
|
|
