A new method of selective enhancement for FGS is proposed

in which the base layer is encoded with an H.264-compliant encoder

while the enhancement layer

by using the outcome of mode decisions within the base layer

is encoded after performing unequal enhancement on motion compensated sub-blocks of varying partition size.H.264 uses tree-structured motion compensation featuring a flexible hierarchy of partition sizes.In general

a large partition size is appropriate for homogeneous areas of the frame and a small one may be beneficial for motion detailed areas.Analysis shows that

for image sequences with particular characteristics

Region-of-Interest (ROI) may contain a significant part of areas with high motion detail.This feature can be used in FGS coding in which bit planes of macroblocks with small partition sizes are bit-shifted up prior to encoding in order to give them a higher priority and a higher probability of being included in a truncated bit stream

hence maximizing the quality of the potential ROI in the decoded sequence without the need to determine coordinates of the ROI in advance.