Background Clustering Pre-training for Few-shot Segmentation
This addresses a specific bottleneck in few-shot segmentation for computer vision applications, representing an incremental improvement over existing pre-training schemes.
The paper tackles the merged background problem in few-shot segmentation pre-training, where only base classes are labeled as foregrounds, making it hard to distinguish novel classes from actual background. The proposed Background Clustering Pre-Training (BCPT) method achieves advanced performance on PASCAL-5i and COCO-20i benchmarks.
Recent few-shot segmentation (FSS) methods introduce an extra pre-training stage before meta-training to obtain a stronger backbone, which has become a standard step in few-shot learning. Despite the effectiveness, current pre-training scheme suffers from the merged background problem: only base classes are labelled as foregrounds, making it hard to distinguish between novel classes and actual background. In this paper, we propose a new pre-training scheme for FSS via decoupling the novel classes from background, called Background Clustering Pre-Training (BCPT). Specifically, we adopt online clustering to the pixel embeddings of merged background to explore the underlying semantic structures, bridging the gap between pre-training and adaptation to novel classes. Given the clustering results, we further propose the background mining loss and leverage base classes to guide the clustering process, improving the quality and stability of clustering results. Experiments on PASCAL-5i and COCO-20i show that BCPT yields advanced performance. Code will be available.