基于改进CycleGAN的水下桥墩裂缝图像生成

吕振鸣,董绍江,何婧瑶,杨金龙,张佳伟

Underwater bridge pier crack image generation based on improved CycleGAN

Zhenming LV,Shaojiang DONG,Jingyao HE,Jinlong YANG,Jiawei ZHANG

表 4 生成样本和真实样本的目标检测效果验证

Tab.4 Validation of target detection effect for generated and real samples

数据 模型 Precision/% Recall/% F1/% mAP/% 生成样本 YOLOv8n 88.7 87.5 88.1 86.2 RT-DETRr18 88.9 88.1 88.5 86.9 真实样本 YOLOv8n 89.2 86.6 88.0 85.4 RT-DETRr18 89.3 87.8 88.5 86.8