Rethinking Global Context in Crowd Counting

Guolei Sun

doi:10.1007/s11633-023-1475-z

Rethinking Global Context in Crowd Counting

2024-04-15

3

PRE

AI

DOI：10.1007/s11633-023-1475-z

原文链接

原文求助

摘要

En 中文

This paper investigates the role of global context for crowd counting. Specifically, a pure transformer is used to extract features with global information from overlapping image patches. Inspired by classification, we add a context token to the input sequence, to facilitate information exchange with tokens corresponding to image patches throughout transformer layers. Due to the fact that transformers do not explicitly model the tried-and-true channel-wise interactions, we propose a token-attention module (TAM) to recalibrate encoded features through channel-wise attention informed by the context token. Beyond that, it is adopted to predict the total person count of the image through regression-token module (RTM). Extensive experiments on various datasets, including ShanghaiTech, UCF-QNRF, JHU-CROWD++ and NWPU, demonstrate that the proposed context extraction techniques can significantly improve the performance over the baselines.

Keyword:

Crowd counting

vision transformer

global context

attention

density map

Rethinking Global Context in Crowd Counting

摘要

期刊

机构