Collaborative: Compressed Domain Search for Text and Images by Sorted Contexts

West Virginia University Research Corporation, Morgantown WV

Investigators

Abstract

With the dramatic increase in the amount of multimedia data, efficiency considerations make it important to consider ways to keep the data in the compressed form for as much as possible, even when it is searched. We propose to develop methodologies for searching directly on compressed text and images, when the compression is based on the family of compression algorithms that depend on sorted context. We also propose to develop search-aware data compression schemes for text and images based on sorted context that will support compressed-domain search directly on the compressed data. For text, we will explore the applicability of the classical linear exact pattern matching algorithms and binary search algorithms for text compressed with the Burrows-Wheeler transform. We will develop similar compression schemes for images based on sorted image contexts, called LICSIC (Lossless Image Compression by Sorted Image Contexts). The results of the proposed research will be important in environments that require the organization and management of huge collections of text or images. Internet search engines, digital libraries, satellite and space data , geographical information systems, bio-informatics and medical diagnostics are typical candidate applications.