Rate-storage regions for Extractable Source Coding with side information

• Published in: Physical communication Vol. 37; pp. 100845 - 24
• Main Authors: Dupraz, Elsa, Roumy, Aline, Maugey, Thomas, Kieffer, Michel
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01-12-2019; Elsevier
• Subjects: Computer Science; Engineering Sciences; Gaussian sources; General sources; Information Theory; Lossless source coding; Lossy source coding; Signal and Image processing; Source coding with side information; Storage-transmission rate region; Source coding with side information; General sources; Gaussian sources; Lossy source coding; Lossless source coding; Storage-transmission rate region; Information theory
• ISSN: 1874-4907; 1876-3219
• DOI: 10.1016/j.phycom.2019.100845

This papers considers the coding of a source X with J decoders, each having access to a different side information Y(j). We define a new source coding problem called Extractable Source Coding with Side Information (ESC-SI). In this problem, the server stores one single coded description of the source, from which J descriptions can be extracted without re-encoding, depending on the side information available at the decoder. We want to minimize both the storage rate of the source on the server, and the J transmission rates from the server to the decoders. We provide the achievable storage-transmission rate regions for lossless source coding of general, non i.i.d., non-ergodic sources, and the achievable storage-transmission rate-distortion regions for lossy source coding for non i.i.d. Gaussian sources. The regions obtained for such general source models provide insightful design guidelines for practical applications.