LLaFS: When Large Language Models Meet Few-Shot Segmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhu, Lanyun, Chen, Tianrun, Ji, Deyi, Ye, Jieping, Liu, Jun
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909158747406336
author Zhu, Lanyun
Chen, Tianrun
Ji, Deyi
Ye, Jieping
Liu, Jun
author_facet Zhu, Lanyun
Chen, Tianrun
Ji, Deyi
Ye, Jieping
Liu, Jun
contents This paper proposes LLaFS, the first attempt to leverage large language models (LLMs) in few-shot segmentation. In contrast to the conventional few-shot segmentation methods that only rely on the limited and biased information from the annotated support images, LLaFS leverages the vast prior knowledge gained by LLM as an effective supplement and directly uses the LLM to segment images in a few-shot manner. To enable the text-based LLM to handle image-related tasks, we carefully design an input instruction that allows the LLM to produce segmentation results represented as polygons, and propose a region-attribute table to simulate the human visual mechanism and provide multi-modal guidance. We also synthesize pseudo samples and use curriculum learning for pretraining to augment data and achieve better optimization. LLaFS achieves state-of-the-art results on multiple datasets, showing the potential of using LLMs for few-shot computer vision tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2311_16926
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle LLaFS: When Large Language Models Meet Few-Shot Segmentation
Zhu, Lanyun
Chen, Tianrun
Ji, Deyi
Ye, Jieping
Liu, Jun
Computer Vision and Pattern Recognition
This paper proposes LLaFS, the first attempt to leverage large language models (LLMs) in few-shot segmentation. In contrast to the conventional few-shot segmentation methods that only rely on the limited and biased information from the annotated support images, LLaFS leverages the vast prior knowledge gained by LLM as an effective supplement and directly uses the LLM to segment images in a few-shot manner. To enable the text-based LLM to handle image-related tasks, we carefully design an input instruction that allows the LLM to produce segmentation results represented as polygons, and propose a region-attribute table to simulate the human visual mechanism and provide multi-modal guidance. We also synthesize pseudo samples and use curriculum learning for pretraining to augment data and achieve better optimization. LLaFS achieves state-of-the-art results on multiple datasets, showing the potential of using LLMs for few-shot computer vision tasks.
title LLaFS: When Large Language Models Meet Few-Shot Segmentation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2311.16926