New adaptive and progressive image transmission approach using function superpositions

摘要

Abstract. We present a novel approach to adaptive and progressivenimage transmission, based on the decomposition of an image into com-npositions and superpositions of monovariate functions. The monovariatenfunctions are iteratively constructed and transmitted, one after the other,nto progressively reconstruct the original image: the progressive transmis-nsion is performed directly in the 1D space of the monovariate functionsnand independently of any statistical properties of the image. Each mono-nvariate function contains only a fraction of the pixels of the image. Eachnnew transmitted monovariate function adds data to the previously trans-nmitted monovariate functions. After each transmission step, by using thenupdated monovariate functions the image is reconstructed with an in-ncreased resolution. Finally, once all the monovariate functions have beenntransmitted, the original image is reconstructed exactly. This approach isncharacterized by its flexibility and robustness to packet loss: any num-nbers of intermediate transmissions and reconstructions are possible, andnin case of packet loss, the global appearance of the transmitted image isnpreserved. Moreover, the intermediate images can be reconstructed atnany resolution, and for any number of intermediate reconstructions, thenoriginal image will be exactly reconstructed. Finally, the quantity of datanto be transmitted only depends on the image size and is independent ofnthe number of intermediate reconstructions. Our main contributions arenthe modification of the decomposition scheme defined by the Kolmog-norov superposition theorem to enable multiresolution image reconstruc-ntions and its application for progressive image transmission, using suc-ncessively increasing resolutions. We illustrate this approach on severalnimages and evaluate the reconstruction quality, decomposition flexibility,nand error resilience during transmission. © 2010 Society of Photo-Optical Instru-nmentation Engineers. u0001DOI: 10.1117/1.3485757