Different finetune strategy: In Figure 1, we visualize the mIoU curve of different fine-tuning strategies. It can be seen that both layer-based and convolution-based fine-tuning methods bring over-fitting problems. This result shows that traditional fine-tuning methods are not suitable for few-shot segmentation tasks. Directly fine-tuning theparameters ofbackbone infew-shot learning affects the robustness ofFSS models. Therefore, we propose anovelfine-tuning strategy,namely SVF.

Previous works focus their efforts on exploring the support information while paying less attention to the mining of the critical query branch.

In this paper, we focus on utilizing pixel-wise relationships between support and query images to facilitate the few-shot segmentation task.

Few-shot segmentation~(FSS) aims at performing semantic segmentation on novel classes given a few annotated support samples. With a rethink of recent advances, we find that the current FSS framework has deviated far from the supervised segmentation framework: Given the deep features, FSS methods typically use an intricate decoder to perform sophisticated pixel-wise matching, while the supervised segmentation methods use a simple linear classification head. Due to the intricacy of the decoder and its matching pipeline, it is not easy to follow such an FSS framework. This paper revives the straightforward framework of ``feature extractor $+$ linear classification head'' and proposes a novel Feature-Proxy Transformer (FPTrans) method, in which the ``proxy'' is the vector representing a semantic class in the linear classification head. FPTrans has two keypoints for learning discriminative features and representative proxies: 1) To better utilize the limited support samples, the feature extractor makes the query interact with the support features from bottom to top layers using a novel prompting strategy.