TCFormer: Visual Recognition via Token Clustering Transformer

Transformers are widely used in computer vision areas and have achieved remarkable success. Most state-of-the-art approaches split images into regular grids and represent each grid region with a vision token. However, fixed token distribution disregards the semantic meaning of different image region...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-07
Main Authors: Wang, Zeng, Sheng, Jin, Xu, Lumin, Liu, Wentao, Chen, Qian, Ouyang, Wanli, Luo, Ping, Wang, Xiaogang
Format: Article
Language:English
Subjects:
Clustering
Computer vision
Image classification
Image segmentation
Object recognition
Pose estimation
Semantic segmentation
Transformers
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Transformers are widely used in computer vision areas and have achieved remarkable success. Most state-of-the-art approaches split images into regular grids and represent each grid region with a vision token. However, fixed token distribution disregards the semantic meaning of different image regions, resulting in sub-optimal performance. To address this issue, we propose the Token Clustering Transformer (TCFormer), which generates dynamic vision tokens based on semantic meaning. Our dynamic tokens possess two crucial characteristics: (1) Representing image regions with similar semantic meanings using the same vision token, even if those regions are not adjacent, and (2) concentrating on regions with valuable details and represent them using fine tokens. Through extensive experimentation across various applications, including image classification, human pose estimation, semantic segmentation, and object detection, we demonstrate the effectiveness of our TCFormer. The code and models for this work are available at https://github.com/zengwang430521/TCFormer.
ISSN:2331-8422