Balanced Data Augmentation of Object Detection Via Boot-strapping

Sungmin Cho, Jinwook Paeng, Junseok Kwon

Abstract: In this paper, we propose the balanced data augmentation method for object detection via boot-strapping. We demonstrate that the proposed method is a kind of boot-strapping algorithms for object detection and improves the performance of object detection in the VOC dataset. Our method not only makes data balance but also improves the detection accuracy.