Scholarlink

For obstacle detection in road scenes, it is very challenging to detect novel objects that are not seen or barely seen during training. To address this issue, we propose an efficient pipeline for obstacle detection in road scenes based on large-scale unlabeled data. Specifically, we use large-scale unlabeled data to train a closed-set model and a open-set model separately in a pseudo-supervised learning manner, and then iteratively improve the performance of both models through the proposed loop-optimization strategy, which employs some useful tricks to remove false positive detections about corner cases. Experimental evidence demonstrates that our approach achieves new state-of-the-art on the popular CODA dataset.

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

Loop Mining Large-Scale Unlabeled Data for Corner Case Detection in Autonomous Driving

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jiawei Zhao,
  • Yiting Duan,
  • Jinming Su,
  • Wangwang Yang,
  • Tingyi Guo,
  • Xingyue Chen,
  • Junfeng Luo

摘要

For obstacle detection in road scenes, it is very challenging to detect novel objects that are not seen or barely seen during training. To address this issue, we propose an efficient pipeline for obstacle detection in road scenes based on large-scale unlabeled data. Specifically, we use large-scale unlabeled data to train a closed-set model and a open-set model separately in a pseudo-supervised learning manner, and then iteratively improve the performance of both models through the proposed loop-optimization strategy, which employs some useful tricks to remove false positive detections about corner cases. Experimental evidence demonstrates that our approach achieves new state-of-the-art on the popular CODA dataset.