Efficient detection of eye blinking and yawn based on facial video utilizing IPPG technique

摘要

For purpose of efficient detection of eye blinking and yawn of human beings, a facial expressions detection method based upon Imaging Photoplethysmography (IPPG) technique was proposed. In the proposed method, facial videos were transformed into R/G/B signals through spatial pixels averaging on related regions of interest (ROI). Then, the signs of eye blinking and yawn were extracted from G-channel signals that contain more stable eye blinking and yawn components by adaptive energy thresholds. The experiments carried out on eight participants demonstrated that the proposed method had an impressive accuracy in detection. It might provide an efficient and succinct approach for applications of facial expression detection, such as fatigue detection, etc.