GRSDet: Learning to Generate Local Reverse Samples for Few-shot Object Detection

Hefei Mei, Taijin Zhao, Shiyuan Tang, Heqian Qiu, Lanxiao Wang, Minjian Zhang, Fanman Meng, Hongliang Li

Abstract: Highlights•Few-shot object detection based on knowledge transfer easily results in confusion between the novel class and other similar categories.•When directly augmenting the sample of novel classes, it is easy to exacerbate the aliasing of feature distributions.•We propose two plug-and-play models to generate and optimize the low-confused samples without extra networks and datasets.

External IDs:dblp:journals/ijon/MeiZTQWZML25