Density Estimation and Crowd Counting

TL;DR

The paper tackles real-time crowd density estimation in videos by extending diffusion-based density modeling to temporal data and introducing an event-driven sampling strategy based on Farneback optical flow to focus computation on informative frames. Building on diffusion frameworks like CrowdDiff, it processes frames at a reduced resolution ($256\\times256$), uses narrow Gaussian kernels and multiple density realizations, and employs a regression branch with a similarity-based consolidation to produce robust density maps, augmented by an edge-detection overlay to mitigate detail loss. Evaluation on ShanghaiTech demonstrates competitive MAE scores (e.g., 54.7 on Part A and 6.4 on Part B) and substantial frame-efficiency gains (up to 80% frame reduction), underscoring the method’s suitability for real-time monitoring in public safety and event management. Overall, the approach combines diffusion-based density estimation with temporal-aware sampling to deliver accurate, scalable crowd counting in video settings.

Abstract

This study enhances a crowd density estimation algorithm originally designed for image-based analysis by adapting it for video-based scenarios. The proposed method integrates a denoising probabilistic model that utilizes diffusion processes to generate high-quality crowd density maps. To improve accuracy, narrow Gaussian kernels are employed, and multiple density map outputs are generated. A regression branch is incorporated into the model for precise feature extraction, while a consolidation mechanism combines these maps based on similarity scores to produce a robust final result. An event-driven sampling technique, utilizing the Farneback optical flow algorithm, is introduced to selectively capture frames showing significant crowd movements, reducing computational load and storage by focusing on critical crowd dynamics. Through qualitative and quantitative evaluations, including overlay plots and Mean Absolute Error (MAE), the model demonstrates its ability to effectively capture crowd dynamics in both dense and sparse settings. The efficiency of the sampling method is further assessed, showcasing its capability to decrease frame counts while maintaining essential crowd events. By addressing the temporal challenges unique to video analysis, this work offers a scalable and efficient framework for real-time crowd monitoring in applications such as public safety, disaster response, and event management.

Figures (6)

• Figure 1: Flow of paper's approach
• Figure 2: Combining Edge Detection
• Figure 3: Video Sampling Approach
• Figure 4: End-to-end proposed approach
• Figure 5: Part A Samples