Review on Biologic Information Extraction Based on Computer Technology

摘要

With the growing development of computer technology and sensor technology, artificial intelligence has been applied to a variety of scenarios, providing solutions to problems in different industries. It is a new dimension in the research of machine vision artificial intelligence to perceive and recognize human biological information, such as physiological state, emotion and biological identity, which contributes to making auxiliary judgment and decision. Based on the extraction of human body-related biological information, this paper aims to review on three aspects which are image-based physiological signal acquisition, sleep information extraction and image-based emotion recognition. With respect to image-based physiological signal acquisition, blind source separation and Eulerian video magnification technology are two more commonly used methods for non-contact video to extract pulse information. Based on the principle of photo plethysmography (PPG), there are also many other methods such as signal weighting analysis and supervised learning algorithm, which are gradually applied to non-contact video acquisition of pulse-related physiological signals. In aspect of sleep information extraction, sleep quality can be analyzed through heart rate and breathing signals. When it comes to image-based emotion recognition, there is still some room for improvement in the accuracy of parameter extraction.