A Neuroinspired Contrast Mechanism enables Few-Shot Object Detection

Lingxiao Yang, Dapeng Chen, Yifei Chen, Wei Peng, Xiaohua Xie

Abstract: Highlights•We identified comparison mechanisms as the bottleneck in few-shot object detection.•We proposed two contrast blocks for training and testing, unlike previous approaches.•We based on our blocks to create a new state-of-the-art few-shot object detector.