ZIP: Scalable Crowd Counting via Zero-Inflated Poisson Modeling

TL;DR

This paper introduces ZIP, a scalable crowd counting method that models sparse, discrete counts using Zero-Inflated Poisson likelihood, effectively handling excess zeros and improving accuracy over traditional MSE-based approaches.

Contribution

ZIP is the first scalable crowd counting framework that explicitly models zero-inflation and count discreteness, outperforming existing methods across various model sizes.

Findings

ZIP outperforms state-of-the-art methods on multiple benchmarks.

Theoretical analysis shows a tighter risk bound for ZIP.

ZIP maintains superior performance across different model scales.

Abstract

Most crowd counting methods directly regress blockwise density maps using Mean Squared Error (MSE) losses. This practice has two key limitations: (1) it fails to account for the extreme spatial sparsity of annotations - over 95% of 8x8 blocks are empty across standard benchmarks, so supervision signals in informative regions are diluted by the predominant zeros; (2) MSE corresponds to a Gaussian error model that poorly matches discrete, non-negative count data. To address these issues, we introduce ZIP, a scalable crowd counting framework that models blockwise counts with a Zero-Inflated Poisson likelihood: a zero-inflation term learns the probability a block is structurally empty (handling excess zeros), while the Poisson component captures expected counts when people are present (respecting discreteness). We provide a generalization analysis showing a tighter risk bound for ZIP than MSE-based losses and DMCount provided that the training resolution is moderately large. To assess the scalability of ZIP, we instantiate it on backbones spanning over 100x in parameters/compute. Experiments on ShanghaiTech A & B, UCF-QNRF, and NWPU-Crowd demonstrate that ZIP consistently surpasses state-of-the-art methods across all model scales.