Table 1 Comparative experimental results of various networks
From: Automatic damage detection and localization of ancient city walls—a case study of the Great Wall
Algorithm | mAP@0.5 (%) | GFLOPs | Params/M | FPS |
|---|---|---|---|---|
Faster-RCNN | 71.1 | 841.0 | 33.05 | 10.9 |
YOLOv8s | 69.9 | 28.4 | 11.13 | 270.2 |
YOLOv5x | 73.9 | 203.8 | 86.17 | 52 |
YOLOv5s | 74.4 | 15.8 | 7.01 | 379.8 |
YOLOv5n | 73.6 | 4.1 | 1.76 | 792.2 |
Improved-YOLOv5n | 74.5 | 4.1 | 1.78 | 768.1 |
