Beyond pure quality: Progressive modes, region of interest coding, and real time video decoding for PDE-based image compression

Pascal Peter, Christian Schmaltz, Nicolas Mach, Markus Mainberger, Joachim Weickert
2015 Journal of Visual Communication and Image Representation  
Compared to well-established transform-based image compression methods such as JPEG and JPEG 2000, approaches based on partialdifferential equations (PDEs) are still in a proof-of-concept stage. Nevertheless, they have already surpassed JPEG 2000 quality-wise for medium to high compression rates for certain classes of images. This particularly holds for R-EED, a codec employing edge-enhancing anisotropic diffusion (EED) and rectangular subdivision. However, today's requirements for practically
more » ... iable compression algorithms go beyond pure compression performance. Codecs must also fulfill the individual feature requirements of specific applications such as internet media or medical imaging. In this paper we propose three such features for the R-EED codec. By reordering grey values and exploiting the subdivision scheme, we incorporate a progressive mode into R-EED. In experiments we can verify that this extension outperforms JPEG and JPEG 2000. Additionally, we show that the rectangular subdivision is well-suited for the incorporation of region of interest (ROI) coding. Our ROI-extensions allows to store image parts with different quality depending on their importance. Finally, to demonstrate that R-EED-based decoding can be performed efficiently, we propose a real-time video player that uses R-EED. All of these extensions are compatible with each other and can be used simultaneously.
doi:10.1016/j.jvcir.2015.06.017 fatcat:l4cwjsqgxfff7n6sml2zntdid4