Thermal-Only Crowd Counting with Deployment-Time Privacy Protection

TL;DR

This paper introduces a thermal-only crowd counting framework that enhances privacy by removing RGB data during inference, using depth-to-RGB diffusion models to improve thermal feature extraction, and achieves competitive results without RGB input.

Contribution

The authors propose the first thermal-only crowd counting method that mitigates privacy issues and leverages depth-to-RGB diffusion models for improved thermal feature representation.

Findings

Achieves competitive performance with state-of-the-art RGB-T methods.

Eliminates the need for RGB data during inference, enhancing privacy.

Single-step LCM denoising preserves structural content better than multi-step approaches.

Abstract

While RGB-Thermal crowd counting has shown promise, the paradigm faces critical limitations: RGB data raises privacy concerns in public surveillance, and multi-modal misalignment degrades fusion performance. We propose the first thermal-only framework specifically designed for privacy-conscious crowd counting, eliminating RGB dependency at inference time and substantially reducing the privacy exposure associated with continuous RGB capture in public surveillance deployments. To mitigate thermal ambiguity, we leverage depth-to-RGB diffusion models as a cross-modal bridge, extracting discriminative features that enhance thermal representations. Critically, we demonstrate that single-step LCM denoising yields features most faithful to the structural content of the depth conditioning signal, while multi-step approaches progressively decouple features from the conditioning input and accumulate errors that degrade counting accuracy. Experiments on RGBT-CC and DroneRGBT datasets show our method achieves competitive performance against state-of-the-art RGB-T fusion methods, while requiring only thermal input during inference, eliminating the need for continuous RGB capture that constitutes the primary privacy concern in real-world surveillance deployment. The code will be made publicly available.