DSV-LFS: Unifying LLM-Driven Semantic Cues with Visual Features for Robust Few-Shot Segmentation

The paper 'DSV-LFS: Unifying LLM-Driven Semantic Cues with Visual Features for Robust Few-Shot Segmentation' by Amin Karimi and Charalambos Poullis has been accepted for publication in IEEE/CVF Computer Vision and Pattern Recognition (CVPR) 2025.

TL;DR: The paper introduces DSV-LFS, a framework that boosts few-shot semantic segmentation by combining a multimodal large language model to generate semantic prompts with a dense pixel-wise matching module for visual prompts, achieving state-of-the-art performance on challenging benchmarks.

Research paper & Supplementary material: https://arxiv.org/abs/2503.04006

Source code: https://github.com/aminpdik/DSV-LFS