Table 2 Performance comparison of different YOLO versions with YOLOv10-ST (proposed) on image and video-based indian sign language recognition.
Model | Data | P | R | F1 score |
|---|---|---|---|---|
YOLOv3 | Image | 84.34 | 86.21 | 84.92 |
Video | 82.45 | 84.13 | 81.87 | |
YOLOv4 | Image | 86.67 | 88.45 | 86.13 |
Video | 85.78 | 86.56 | 84.92 | |
YOLOv5 | Image | 89.13 | 90.78 | 88.45 |
Video | 87.12 | 89.54 | 86.56 | |
YOLOv6 | Image | 91.25 | 91.34 | 90.67 |
Video | 89.56 | 90.12 | 88.89 | |
YOLOv7 | Image | 93.89 | 93.76 | 91.23 |
Video | 90.32 | 92.45 | 89.76 | |
YOLOv8 | Image | 94.78 | 94.12 | 92.02 |
Video | 91.78 | 93.01 | 91.45 | |
YOLOv9 | Image | 95.76 | 96.13 | 93.54 |
Video | 92.25 | 94.78 | 92.01 | |
YOLOv10 | Image | 96.87 | 97.34 | 94.13 |
Video | 94.17 | 95.27 | 93.32 | |
YOLOv10-ST (Proposed) | Image | 97.5 | 98.1 | 96.58 |
Video | 95.24 | 96.0 | 95.87 |
