A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf
.
Effect of Annotation Errors on Drone Detection with YOLOv3
[article]
2020
arXiv
pre-print
Following the recent advances in deep networks, object detection and tracking algorithms with deep learning backbones have been improved significantly; however, this rapid development resulted in the necessity of large amounts of annotated labels. Even if the details of such semi-automatic annotation processes for most of these datasets are not known precisely, especially for the video annotations, some automated labeling processes are usually employed. Unfortunately, such approaches might
arXiv:2004.01059v3
fatcat:wyetbg7kijfplnjmtqqljhxz3u