Weide Liu, Zhonghua Wu, Yang Zhao, Yuming Fang, Chuan-Sheng Foo, Jun Cheng, Guosheng Lin
Current methods for few-shot segmentation (FSSeg) have mainly focused on improving the performance of novel classes while neglecting the performance of base classes. To overcome this limitation, the task of generalized few-shot semantic segmentation (GFSSeg) has been introduced, aiming to predict segmentation masks for both base and novel classes. However, the current prototype-based methods do not explicitly consider the relationship between base and novel classes when updating prototypes, leading to a limited performance in identifying true categories. To address this challenge, we propose a class contrastive loss and a class relationship loss to regulate prototype updates and encourage a large distance between prototypes from different classes, thus distinguishing the classes from each other while maintaining the performance of the base classes. Our proposed approach achieves new state-of-the-art performance for the generalized few-shot segmentation task on PASCAL VOC and MS COCO datasets.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Few-Shot Learning | COCO-20i (1-shot) | Mean Base and Novel | 27.86 | CCA (ResNet-50) |
| Few-Shot Learning | COCO-20i (1-shot) | Mean IoU | 37.48 | CCA (ResNet-50) |
| Few-Shot Semantic Segmentation | COCO-20i (1-shot) | Mean Base and Novel | 27.86 | CCA (ResNet-50) |
| Few-Shot Semantic Segmentation | COCO-20i (1-shot) | Mean IoU | 37.48 | CCA (ResNet-50) |
| Meta-Learning | COCO-20i (1-shot) | Mean Base and Novel | 27.86 | CCA (ResNet-50) |
| Meta-Learning | COCO-20i (1-shot) | Mean IoU | 37.48 | CCA (ResNet-50) |