In computer vision, all we do depends on the photos or video material we analyse. It guides even the most powerful neural network that encodes the largest amount of prior knowledge. In this talk, I will shed some light on how illumination, movement, camera specifications, and video encoding affect video quality. My students have quantified the VIIDEO score for our Parkinson's disease and rPPG datasets, which we use to track skin features and quantify the heart rate.
Torbjörn Nordling, Principal Investigator at Nordling Lab, Department of Mechanical Engineering at the National Cheng Kung University in Tainan, Taiwan. Associate professor (on leave) at Department of Applied Physics and Electronics, Umeå University.