Crowd Counting via Wi-Fi Probe Requests: Integrating Feature Selection and Data Generation

Crowd monitoring is essential for smart city applications, particularly for optimizing public transit systems. To address this need, we propose a privacy-conscious crowd counting pipeline using Wi-Fi probe requests, designed to adapt to the challenges posed by MAC address randomization. Our approach leverages a random forest-based feature selection process to identify key Information Elements and frame attributes, and applies DBSCAN clustering with adaptive parameter optimization for device counting. To mitigate the limited availability of labeled data, a diffusion model generates synthetic tabular data, enhancing model robustness. Experimental results demonstrate improved accuracy in device counting, achieving a V-measure of 0.952, an average silhouette score of 0.789, and reliable clustering counts.