Generalized Few-Shot Semantic Segmentation: All You Need is Fine-Tuning
This work addresses the challenge of improving segmentation models for both novel and base classes in few-shot learning, offering a practical solution for applications in computer vision, though it is incremental as it builds on fine-tuning techniques.
The paper tackled the problem of generalized few-shot semantic segmentation, which requires models to segment both novel and base classes, by proposing a fine-tuning solution that addresses saturation issues in meta-learning approaches. The method achieved state-of-the-art results on PASCAL-5i and COCO-20i datasets, outperforming existing methods with concrete performance gains.
Generalized few-shot semantic segmentation was introduced to move beyond only evaluating few-shot segmentation models on novel classes to include testing their ability to remember base classes. While the current state-of-the-art approach is based on meta-learning, it performs poorly and saturates in learning after observing only a few shots. We propose the first fine-tuning solution, and demonstrate that it addresses the saturation problem while achieving state-of-the-art results on two datasets, PASCAL-5i and COCO-20i. We also show that it outperforms existing methods, whether fine-tuning multiple final layers or only the final layer. Finally, we present a triplet loss regularization that shows how to redistribute the balance of performance between novel and base categories so that there is a smaller gap between them.