Optical EngineeringComparison of the predictions of a spatio temporal model with the detection of distortion in small moving images
|Format||Member Price||Non-Member Price|
The image sequence discrimination model we use models optical blurring and retinal light adaptation. Further, two parallel channels, sustained and transient, with different masking rules based on contrast gain control, are also used. Performance of the model is studied for two tasks representative of a video communication system with versions of monochrome H.263 compressed images. In the first study, five image sequences constitute pairs of noncompressed and compressed images to be discriminated with a two-alternative-forced-choice method together with a staircase procedure. The discrimination thresholds for each subject are calculated. Analysis of variance shows that the differences between the pictures are significant. The model threshold is close to the average threshold of the subjects for each picture, and the model thus predicts these results quite well. In the second study, the effect of transmission errors on the Internet, i.e., packet losses, is tested with the method of constant stimuli. Both reference and comparison images are distorted. The task of the subjects is to judge whether the presented video quality is worse than the initially seen reference video. Two different quality levels of the compressed sequences are simulated. The differences in the thresholds among the different video scenes are to some extent predicted by the model. Category scales indicate that detection of distortions and overall quality judgements are based on different psychological processes.