Probabilistic Prototype Calibration of Vision-Language Models for Generalized Few-shot Semantic Segmentation

TL;DR

FewCLIP introduces a probabilistic calibration framework for vision-language models that enhances adaptability and generalization in generalized few-shot semantic segmentation, outperforming existing methods on standard benchmarks.

Contribution

It proposes a novel probabilistic prototype calibration method for GFSS using CLIP, improving adaptability and reducing overfitting in few-shot scenarios.

Findings

Outperforms state-of-the-art on PASCAL-5i and COCO-20i datasets.

Provides uncertainty-aware prototype learning for better generalization.

Demonstrates effectiveness in class-incremental settings.

Abstract

Generalized Few-Shot Semantic Segmentation (GFSS) aims to extend a segmentation model to novel classes with only a few annotated examples while maintaining performance on base classes. Recently, pretrained vision-language models (VLMs) such as CLIP have been leveraged in GFSS to improve generalization on novel classes through multi-modal prototypes learning. However, existing prototype-based methods are inherently deterministic, limiting the adaptability of learned prototypes to diverse samples, particularly for novel classes with scarce annotations. To address this, we propose FewCLIP, a probabilistic prototype calibration framework over multi-modal prototypes from the pretrained CLIP, thus providing more adaptive prototype learning for GFSS. Specifically, FewCLIP first introduces a prototype calibration mechanism, which refines frozen textual prototypes with learnable visual calibration prototypes, leading to a more discriminative and adaptive representation. Furthermore, unlike deterministic prototype learning techniques, FewCLIP introduces distribution regularization over these calibration prototypes. This probabilistic formulation ensures structured and uncertainty-aware prototype learning, effectively mitigating overfitting to limited novel class data while enhancing generalization. Extensive experimental results on PASCAL-5$^i$ and COCO-20$^i$ datasets demonstrate that our proposed FewCLIP significantly outperforms state-of-the-art approaches across both GFSS and class-incremental setting. The code is available at https://github.com/jliu4ai/FewCLIP.