A Phantom Study of Using Moiré Effect to Significantly Enhance Camera-based Respiratory Motion Extraction

Camera-based respiration monitoring has gained popularity in research and application in the last decade due to its non-contact and ubiquitous nature. It typically applies optical flow methods to track subtle respiratory movements in the chest and abdomen region. However, various real-world challenges, including shallow breathing, thick blanket obstruction, and low-textured surface, hinder the sensitivity of respiration monitoring. This study proposes a novel concept of creating moiré patterns to improve the performance of camera-based respiratory signal extraction. The moiré effect is created by the interference between dense fabric texture and the camera sensor grid, which amplifies the pixel displacement caused by respiratory motion. This approach utilizes the inherent properties of fabrics without the need for additional texture attachments or hardware adaptation. Our modeling and experiments show that the moiré effect can significantly enhance the sensitivity of respiratory signal extraction, reducing the errors for respiratory rate estimation while expanding the valid image area for respiratory signal extraction. This approach provides a novel and powerful enhancement to camera-based respiration monitoring systems.