Training;Image quality;Computer vision;Computational modeling;Conferences;Training data;Distortion