Freezing the pre-trained backbone has become a standard paradigm to avoid overfitting in few-shot segmentation. In this paper, we rethink the paradigm and explore a new regime: {\em fine-tuning a small part of parameters in the backbone}. We present a solution to overcome the overfitting problem, leading to better model generalization on learning novel classes. Our method decomposes backbone parameters into three successive matrices via the Singular Value Decomposition (SVD), then {\em only fine-tunes the singular values} and keeps others frozen. The above design allows the model to adjust feature representations on novel classes while maintaining semantic clues within the pre-trained backbone. We evaluate our {\em Singular Value Fine-tuning (SVF)} approach on various few-shot segmentation methods with different backbones. We achieve state-of-the-art results on both Pascal-5i and COCO-20i across 1-shot and 5-shot settings. Hopefully, this simple baseline will encourage researchers to rethink the role of backbone fine-tuning in few-shot settings. The source code and models will be available at https://github.com/syp2ysy/SVF.
@article{arxiv.2206.06122,
title = {Singular Value Fine-tuning: Few-shot Segmentation requires Few-parameters Fine-tuning},
author = {Yanpeng Sun and Qiang Chen and Xiangyu He and Jian Wang and Haocheng Feng and Junyu Han and Errui Ding and Jian Cheng and Zechao Li and Jingdong Wang},
journal= {arXiv preprint arXiv:2206.06122},
year = {2022}
}