University of Victoria (Canada)
Adaptive strategies and optimization techniques for JPEG-based low bit-rate image coding
Abstract
dc:description.abstractThe field of digital image compression has been intensively explored to obtain ever improved performance corresponding to a given bit budget. The DCT-based JPEG standard remains to be one of the most popular image compression standards due to its reasonable coding performance, fast implementations, friendly low-cost architecture. flexibility and adaptivity on block level. In this thesis, we consider the problem of low bit-rate image coding and present new approaches using adaptive strategies and optimization techniques for performance enhancement, while employing the DCT block-based JPEG standard as the main framework with several pre- and post-processing , steps. We propose an adaptive coding approach which involves a variable quality factor in the quantization step of JPEG compression to make the compression more flexible in respect to bit budget requirements. We also propose an adaptive sampling approach based on variable down-'up-scaling rate and local image characteristics. In addition. we study an adaptive filtering approach in which the optimal filters coefficients are determined by making use of optimization methods and symmetric extension techniques. Simulation results are presented to demonstrate the effectiveness of the proposed techniques relative to recent works in the field of low bit-rate image coding.
Degree
thesis:*- Level thesis:degree_level
- Master of Applied Science M.A.Sc.
- Department thesis:degree_department
- Department of Electrical and Computer Engineering
- Year dc:date.issued
- 2010
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Sevcenco, Ana-Maria
- Advisor dc:contributor.supervisor
-
- Lu, Wu-Sheng
Subjects
dc:subject × 3Rights
dc:rights- Statement dc:rights
-
- Available to the World Wide Web
- Language dc:language.iso
- en, English
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- http://hdl.handle.net/1828/2282