Search papers, labs, and topics across Lattice.
This paper addresses the challenge of unbalanced region proposal distributions in few-shot object detection by introducing a proposal refinement approach tailored for different training phases. The method enhances model sensitivity to novel classes during the base training phase through a specially designed refinement loss, and it incorporates an auxiliary refinement branch in the fine-tuning phase to generate additional novel proposals. Experimental results demonstrate that this approach achieves a state-of-the-art performance improvement of 1% to 6% on existing benchmarks without increasing inference time.
By rebalancing proposal distributions, this method sets a new benchmark in few-shot object detection, achieving significant performance gains without added latency.
Few-shot object detection has gained widely attention in recent years. Some excellent algorithms have been proposed to handle this task. However, most of these algorithms rely on the performance of few-shot classification. Unlike previous attempts, our work focuses on the problem of unbalanced distribution of region proposals between the novel classes and the base classes. In order to alleviate this unbalanced distribution, we propose the proposal refinement approach for different training phases. Specifically, refinement loss is designed for the base training phase to enhance sensitivity of the model to novel classes, and refinement branch is introduced as an auxiliary branch for RPN (Region Proposal Networks) to generate more novel proposals in the fine-tuning phase. By rebalancing the proposal distribution, the proposed approach outperforms the baselines methods by roughly 1\%$\sim$6\% on current benchmarks without increasing any inference time. Through extensive experiments, we prove that we establish a new state-of-the-art method for the few-shot object detection task.
Tsinghua AI