JPEG2000 block-wise truncation of quality layers

Quality scalability and quality progression are fundamental features of JPEG2000, achieved through the use of quality layers that are optimally formed in the encoder by rate-distortion optimization techniques. The successive truncation and decoding of the codestream at quality layers boundaries achieves optimal rate-distortion representations of the image, and when the image is interactively transmitted, quality layers supply successive quality refinements for the transmitted windows of interest. Although quality layers are a sound mechanism of JPEG2000, their practical use must consider that codestreams containing an inadequate allocation of layers may greatly penalize the coding performance, especially when few quality layers are available and truncations at any point rather than at quality layers boundaries are carried out. In interactive image transmission, real-time video rendering, and transcoding applications, for instance, this issue represents an unacceptable shortcoming. This paper introduces a block-wise truncation of quality layers that, requiring negligible computational resources, enhances the coding performance of codestreams containing an inadequate layer allocation. Experimental results suggest that the proposed strategy achieves near-optimal results even when the codestream contains a single quality layer.