Dense Scale Network for Crowd Counting

TL;DR

This paper introduces DSNet, a dense scale network for crowd counting that effectively captures multiple scales through dense dilated convolutions and residual connections, significantly improving accuracy across several datasets.

Contribution

The paper proposes a novel dense dilated convolution block with scale-preserving connections and a multi-scale density level consistency loss for enhanced crowd counting performance.

Findings

Achieves state-of-the-art results on four datasets

Improves accuracy by up to 30% over previous methods

Effectively captures a wide range of scales in crowd images

Abstract

Crowd counting has been widely studied by computer vision community in recent years. Due to the large scale variation, it remains to be a challenging task. Previous methods adopt either multi-column CNN or single-column CNN with multiple branches to deal with this problem. However, restricted by the number of columns or branches, these methods can only capture a few different scales and have limited capability. In this paper, we propose a simple but effective network called DSNet for crowd counting, which can be easily trained in an end-to-end fashion. The key component of our network is the dense dilated convolution block, in which each dilation layer is densely connected with the others to preserve information from continuously varied scales. The dilation rates in dilation layers are carefully selected to prevent the block from gridding artifacts. To further enlarge the range of scales covered by the network, we cascade three blocks and link them with dense residual connections. We also introduce a novel multi-scale density level consistency loss for performance improvement. To evaluate our method, we compare it with state-of-the-art algorithms on four crowd counting datasets (ShanghaiTech, UCF-QNRF, UCF_CC_50 and UCSD). Experimental results demonstrate that DSNet can achieve the best performance and make significant improvements on all the four datasets (30% on the UCF-QNRF and UCF_CC_50, and 20% on the others).