Open-Vocabulary SAM: Segment and Recognize Twenty-thousand Classes Interactively

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yuan, Haobo, Li, Xiangtai, Zhou, Chong, Li, Yining, Chen, Kai, Loy, Chen Change
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909315643736064
author Yuan, Haobo
Li, Xiangtai
Zhou, Chong
Li, Yining
Chen, Kai
Loy, Chen Change
author_facet Yuan, Haobo
Li, Xiangtai
Zhou, Chong
Li, Yining
Chen, Kai
Loy, Chen Change
contents The CLIP and Segment Anything Model (SAM) are remarkable vision foundation models (VFMs). SAM excels in segmentation tasks across diverse domains, whereas CLIP is renowned for its zero-shot recognition capabilities. This paper presents an in-depth exploration of integrating these two models into a unified framework. Specifically, we introduce the Open-Vocabulary SAM, a SAM-inspired model designed for simultaneous interactive segmentation and recognition, leveraging two unique knowledge transfer modules: SAM2CLIP and CLIP2SAM. The former adapts SAM's knowledge into the CLIP via distillation and learnable transformer adapters, while the latter transfers CLIP knowledge into SAM, enhancing its recognition capabilities. Extensive experiments on various datasets and detectors show the effectiveness of Open-Vocabulary SAM in both segmentation and recognition tasks, significantly outperforming the naïve baselines of simply combining SAM and CLIP. Furthermore, aided with image classification data training, our method can segment and recognize approximately 22,000 classes.
format Preprint
id arxiv_https___arxiv_org_abs_2401_02955
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Open-Vocabulary SAM: Segment and Recognize Twenty-thousand Classes Interactively
Yuan, Haobo
Li, Xiangtai
Zhou, Chong
Li, Yining
Chen, Kai
Loy, Chen Change
Computer Vision and Pattern Recognition
The CLIP and Segment Anything Model (SAM) are remarkable vision foundation models (VFMs). SAM excels in segmentation tasks across diverse domains, whereas CLIP is renowned for its zero-shot recognition capabilities. This paper presents an in-depth exploration of integrating these two models into a unified framework. Specifically, we introduce the Open-Vocabulary SAM, a SAM-inspired model designed for simultaneous interactive segmentation and recognition, leveraging two unique knowledge transfer modules: SAM2CLIP and CLIP2SAM. The former adapts SAM's knowledge into the CLIP via distillation and learnable transformer adapters, while the latter transfers CLIP knowledge into SAM, enhancing its recognition capabilities. Extensive experiments on various datasets and detectors show the effectiveness of Open-Vocabulary SAM in both segmentation and recognition tasks, significantly outperforming the naïve baselines of simply combining SAM and CLIP. Furthermore, aided with image classification data training, our method can segment and recognize approximately 22,000 classes.
title Open-Vocabulary SAM: Segment and Recognize Twenty-thousand Classes Interactively
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2401.02955