Context-Aware Crowd Counting

TL;DR

This paper presents an end-to-end deep learning model for crowd counting that adaptively combines features from multiple receptive fields, improving accuracy especially in scenes with strong perspective distortions.

Contribution

The proposed method introduces a novel end-to-end trainable architecture that adaptively learns the importance of features from various receptive fields for crowd density estimation.

Findings

Outperforms state-of-the-art methods in crowd counting accuracy.

Effectively handles scenes with strong perspective distortions.

Demonstrates superior adaptability to varying scene scales.

Abstract

State-of-the-art methods for counting people in crowded scenes rely on deep networks to estimate crowd density. They typically use the same filters over the whole image or over large image patches. Only then do they estimate local scale to compensate for perspective distortion. This is typically achieved by training an auxiliary classifier to select, for predefined image patches, the best kernel size among a limited set of choices. As such, these methods are not end-to-end trainable and restricted in the scope of context they can leverage. In this paper, we introduce an end-to-end trainable deep architecture that combines features obtained using multiple receptive field sizes and learns the importance of each such feature at each image location. In other words, our approach adaptively encodes the scale of the contextual information required to accurately predict crowd density. This yields an algorithm that outperforms state-of-the-art crowd counting methods, especially when perspective effects are strong.