Using subjective redundancy for DCT coding of moving images

Publisher Website
Google Scholar
Abstract
An investigation is carried out on employing subjective redundancy to improve DCT coding respecting the non-trackable motion in video signals. The human visual system (HVS) is exploited through the two aspects of its behavior, the sensitivity to a visual stimulus and visual masking. Some measures possible to adopt those aspects are discussed and tested. The result showed that: (1) during the non-trackable motion, a reduction of the HVSs response to the reconstruction error occurs because of the loss of sensitivity and the increase of the masking effect; (2) as far as the sensitivity is concerned, adjusting the quantization weighting function with respect to the motion is unlikely to bring substantial improvement; (3) the magnitude of the visual masking by the image content over the reconstruction error is positively related to the function of spatial masking and the relative speed between the eyes and the image. An example of application on motion compensated DCT coding is presented.

This publication has 0 references indexed in Scilit: