Perspective-Guided Convolution Networks for Crowd Counting

TL;DR

This paper introduces PGCNet, a novel CNN-based crowd counting method that uses perspective information to adaptively smooth features, improving accuracy over existing multi-scale approaches, and also presents a new large-scale dataset.

Contribution

The paper proposes a perspective-guided convolution approach for crowd counting that models continuous scale variations and introduces a new large-scale crowd dataset.

Findings

PGCNet outperforms state-of-the-art methods on four benchmark datasets.

The perspective estimation branch can be trained in supervised or weakly-supervised modes.

The new dataset contains over 13,000 high-resolution crowd images.

Abstract

In this paper, we propose a novel perspective-guided convolution (PGC) for convolutional neural network (CNN) based crowd counting (i.e. PGCNet), which aims to overcome the dramatic intra-scene scale variations of people due to the perspective effect. While most state-of-the-arts adopt multi-scale or multi-column architectures to address such issue, they generally fail in modeling continuous scale variations since only discrete representative scales are considered. PGCNet, on the other hand, utilizes perspective information to guide the spatially variant smoothing of feature maps before feeding them to the successive convolutions. An effective perspective estimation branch is also introduced to PGCNet, which can be trained in either supervised setting or weakly-supervised setting when the branch has been pre-trained. Our PGCNet is single-column with moderate increase in computation, and extensive experimental results on four benchmark datasets show the improvements of our method against the state-of-the-arts. Additionally, we also introduce Crowd Surveillance, a large scale dataset for crowd counting that contains 13,000+ high-resolution images with challenging scenarios.