Proposal Refinement for Few-Shot Object Detection
2026-06-08 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors focus on improving few-shot object detection by addressing an imbalance problem where the model sees fewer region proposals for new object classes compared to known ones. They create a 'proposal refinement' method that changes how the model trains, making it pay more attention to new classes during both initial training and fine-tuning. This approach improves the model's ability to detect new objects without slowing down the detection process. Their experiments show their method performs better than previous ones by a small but meaningful margin.
few-shot object detectionregion proposalsnovel classesbase classesproposal refinementregion proposal network (RPN)fine-tuningtraining phasesrefinement loss
Authors
Yuan Zeng, Bin Song, Jie Guo, Yuwen Chen
Abstract
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.