The natural progression would be to establish the effect on bit rates of implementing an adaptive DCT method.
A more uniform disparity map such as one employing a segmentation
approach to the
image [9] would yield better compression results and
reduce blocking artefacts.
Another step would be to investigate methods of compressing colour images as well as simply greyscale images. These should yield greater compression rates [6].
A possible approach to better methods is to establish whether one eye is in general dominant in human observers, in the same way that a majority of people are right-handed. If this is the case, then one could bias the compression process in favour of the image seen by that eye.