The Impact of Spatial Masking in Image Quality Meters
Compression of digital image and video leads to block-based visible distortions like blockiness. The PSNR quality metric doesn’t correlate well with the subjective metric as it doesn’t take into consideration the impact of human visual system. In this work, we study the impact of human visual system in masking the coding distortions and its effect on the accuracy of the quality meter. We have chosen blockiness which is the most common coding distortion in DCTbased JPEG or intracoded video. We have studied the role of spatial masking by applying different masking techniques on full, reduced and no reference meters. As the visibility of distortion is content dependent, the distortion needs to be masked according to the spatial activity of the image. The results show that the complexity of spatial masking may be reduced by using the reference information efficiently. For full and reduced reference meters the spatial masking hasn’t much importance, if the blockiness detection is accurate, while for the no reference meter spatial masking is required to compensate the absence of any required reference information.
