Enhancing In-Cabin Video Experience Using Driver Monitoring IR Camera

Safety regulations such as Euro NCAP and the General Safety Regulation (GSR) mandate the inclusion of in-cabin infrared (IR) cameras in every new car for enhanced safety and monitoring. IR cameras are highly effective in detecting faces and extracting facial features, making them an excellent choice for driver monitoring systems. This paper explores the use of these pre-installed IR cameras beyond safety applications, specifically for real-time video conferencing. Leveraging artificial intelligence (AI) techniques, we propose a framework that enhances IR-based video communication while maintaining privacy and robustness under various lighting conditions. The system integrates the Mediapipe framework for face detection, tracking, and segmentation, combined with AI-driven image enhancement techniques to improve the visual quality of IR-based video feeds. To assess the effectiveness of our approach, we evaluate multiple qualitative and quantitative criteria, including sharpness, noise levels, facial feature clarity, edge artifacts, and video smoothness. Our results demonstrate that IR cameras, when enhanced with AI techniques, can serve as a viable alternative to traditional RGB-based in-cabin communication systems, offering a cost-effective and privacy-preserving solution.