SAM-IF: Leveraging SAM for Incremental Few-Shot Instance Segmentation
This addresses incremental few-shot instance segmentation for computer vision applications, but appears incremental as it builds on SAM with classifier modifications.
The paper tackles the problem of class-agnostic instance segmentation in incremental few-shot settings by proposing SAM-IF, which fine-tunes the Segment Anything Model with a multi-class classifier and cosine-similarity-based adaptation. It achieves competitive results for specific object segmentation with limited labeled data.
We propose SAM-IF, a novel method for incremental few-shot instance segmentation leveraging the Segment Anything Model (SAM). SAM-IF addresses the challenges of class-agnostic instance segmentation by introducing a multi-class classifier and fine-tuning SAM to focus on specific target objects. To enhance few-shot learning capabilities, SAM-IF employs a cosine-similarity-based classifier, enabling efficient adaptation to novel classes with minimal data. Additionally, SAM-IF supports incremental learning by updating classifier weights without retraining the decoder. Our method achieves competitive but more reasonable results compared to existing approaches, particularly in scenarios requiring specific object segmentation with limited labeled data.