Scholarlink

To analyze the impact of individual parameters, 100 images were randomly selected, and each state parameter was set to its maximum value. The detection results of the YOLOv5 object recognition network on these images were recorded. By comparing the detected bounding boxes with the ground truth labels, the IoU and confidence of the network were calculated.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Discussion on Reinforcement Learning-Based Scene Data Augmentation

  • 章节

  • 发表日期：

    2026年

查看全文
  • Kun Gao

摘要

To analyze the impact of individual parameters, 100 images were randomly selected, and each state parameter was set to its maximum value. The detection results of the YOLOv5 object recognition network on these images were recorded. By comparing the detected bounding boxes with the ground truth labels, the IoU and confidence of the network were calculated.