WiFi Router Project Tracks Human Poses Through Walls / X

Developer Reuven Cohen released an open-source project called wifi-densepose that uses WiFi signal analysis to detect and track human poses through walls. The system can monitor up to 10 people simultaneously at 30-50 frames per second, building on earlier research from Carnegie Mellon and MIT. Rather than relying on cameras, the technology analyzes how WiFi signals change when passing through and reflecting off human bodies, making it a privacy-preserving alternative for monitoring applications.

The project gained over 18,000 GitHub stars, demonstrating significant developer interest. Cohen has proposed practical applications in healthcare settings and smart homes, where tracking movement patterns could support patient monitoring, fall detection, or occupancy awareness without privacy concerns associated with video surveillance. The technology requires CSI-enabled WiFi hardware like ESP32-S3 devices, making it accessible for implementation.

Initial skepticism about the project's viability has been addressed through technical improvements. Cohen resolved early issues with deployment and provided benchmarks demonstrating the system's actual performance on supported hardware. This open-source approach enables other developers to build on the foundation and adapt the technology for specific use cases. The combination of wall-penetrating sensing and pose tracking creates new possibilities for autonomous monitoring systems that operate without visual cameras.