Researchers found that chrominance signals extracted from the upper chest by cameras estimated respiratory rate more accurately than facial video signals. The untrained chest chrominance signal had a mean absolute error of 1.267 breaths/min, compared with 3.693–3.912 breaths/min for facial color channels. The evaluation used 160 video segments and 39 participants from COHFACE and compared the results with the VitalVideos dataset. The study also recommends further investigation of skin tone, clothing texture, and motion variations in real clinical environments.
Source:Sensor Technology · Read Original
Leave a Reply