GoMap Verification: A Labelled Traffic Sign Source

2022 23rd IEEE International Conference on Mobile Data Management (MDM)(2022)

引用 0|浏览12
暂无评分
摘要
Images are an important data source for many applications such as autonomous driving and map annotations. However, it is expensive to collect large imagery sets and annotate objects such as traffic signs. In the paper, we present a web solution where a user efficiently can label traffic signs by verifying or correcting the labels provided by an object-detection model. All verified/corrected objects can immediately be downloaded by the user for usage in for example semi-supervised or supervised learning. The tool is available via the URL https://gomap.cs.aau.dk
更多
查看译文
关键词
imagery,traffic sign,object detection,open data
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要