In this paper, we investigated the correlation between blood pressure (BP) and image-based pulse transit time (iPTT) acquired from only face using RGB camera. In general, the value of iPTT can be calculated from a transition time at peaks of hemoglobin amount which is extracted from RGB values of the camera. The transition time of peaks is obtained by time of the peak at the face and the palm. In the previous research, it is known that there is a correlation between BP and iPTT. Therefore, blood pressure can be estimated by acquiring iPTT from video taken using RGB camera. However, it is necessary to take video of face and palm simultaneously in this conventional iPTT measurement method. It is difficult to take the video of face and palm simultaneously. In order to solve this problem, we investigated whether iPTT can be acquired from a single part of body. At first, we took a video of face and palm using a high-speed camera and investigated whether the pulse wave flow can be observed. As a result, we were able to observe the propagation of the pulse wave in the regions of face and palm. Hence, iPTT acquisition from the single part of body is expected to be possible in these two parts. Next, we set a region of interest (ROI) on the chin and the forehead for the face, and the center of the palm and thenar for the palm. Then, pulse waves were extracted from each region, and iPTT was calculated from each part. We found that iPTT acquired from face is long and stable, but iPTT acquired from palm is short and unstable. Moreover, we examined the correlation between blood pressure and iPTT acquired from face. Consequently, a correlation was found between blood pressure and iPTT acquired from only face.
Ryo Takahashi, Keiko Ogawa-Ochiai, Norimichi Tsumura, "Improvement of blood pressure estimation from face video using RGB camera" in Proc. IS&T 27th Color and Imaging Conf., 2019, pp 273 - 278, https://doi.org/10.2352/issn.2169-2629.2019.27.49