HKUST Institutional Repository >
Electronic and Computer Engineering >
ECE Master Theses >
Please use this identifier to cite or link to this item:
|Title: ||Very low bit rate fractal video coding by genetic algorithm|
|Authors: ||Fu, King Wa|
|Issue Date: ||1995 |
|Abstract: ||With the increasing demand in multimedia applications through various kinds of communication channels, very low bit rate (VLBR) video coding has become an active research topic. The current video coding standards, such as MPEG-2/H.262 or H.261, are not capable of operating in narrow bandwidth applications, such as telecommunication through analog telephone line or wireless communication. Hence, new techniques, other than conventional methods such as Discrete Cosine Transform (DCT) and bloc matching motion estimation, are needed for this kind of very low k bit rate applications.
Fractal coding has great potential to achieve high compression ratio in very low bit rate applications. Based on the Iterated Function System (IFS) theory used in fractal modeling, fractal coding exploits the spatial/temporal redundancy through the self-transformability between the partitions in the encoding object. However, the usefulness of fractal coding has been hampered by its large computational complexity.
In this thesis, a fast fractal coding scheme using Genetic Algorithm (GA) is pro-posed to reduce the computational complexity. The proposed scheme can achieve an acceptable image quality with relatively small amount of computation. The simulation results are compared with the international standard JPEG and an existing fast fractal coding method. In addition, we propose a hybrid video coding scheme based on the combination of our fast fractal coding scheme and transform coding. Our simulation results show that we have obtained acceptable image quality by using our proposed scheme in very low bit rate applications.|
|Description: ||Thesis (M.Phil.)--Hong Kong University of Science and Technology, 1995|
94 leaves : ill., photos. (some col.) ; 30 cm
HKUST Call Number: Thesis ELEC 1995 Fu
|Appears in Collections:||ECE Master Theses|
Files in This Item:
All items in this Repository are protected by copyright, with all rights reserved.