東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Fast motion vector estimation techni...

Chalidabhongse, Junavit.

Linked to FindBook

Google Book

Amazon

博客來

Fast motion vector estimation techniques for video coding.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Fast motion vector estimation techniques for video coding./
Author:	Chalidabhongse, Junavit.
Description:	102 p.
Notes:	Source: Dissertation Abstracts International, Volume: 57-07, Section: B, page: 4589.
Contained By:	Dissertation Abstracts International57-07B.
Subject:	Engineering, Electronics and Electrical. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=9636325
ISBN:	0591022346

Fast motion vector estimation techniques for video coding.
Chalidabhongse, Junavit.

Fast motion vector estimation techniques for video coding. - 102 p.

Source: Dissertation Abstracts International, Volume: 57-07, Section: B, page: 4589.

Thesis (Ph.D.)--University of Southern California, 1996.

Motion compensated predictive coding is an effective and commonly, used method to reduce the temporal redundancy existing in image sequences. It predicts an image frame of a sequence based on the motion information between frames. Thus, motion vector (MV) estimation is a key factor in efficient video coding. Block-based motion estimation has been widely used due to its simplicity. The full search block matching algorithm (FBMA) can be performed by searching all locations in a given search area and selecting the position which gives the minimum matching error. Despite providing an optimal solution, FBMA requires a large amount of computation. The high computational complexity of MV estimation is known to be the main bottleneck in real-time encoding.

ISBN: 0591022346Subjects--Topical Terms:

626636
Engineering, Electronics and Electrical.

Fast motion vector estimation techniques for video coding.
LDR:03158nmm 2200289 4500 001 1840157
005 20050714101548.5
008 130614s1996 eng d
020 $a 0591022346
035 $a (UnM)AAI9636325
035 $a AAI9636325
040 $a UnM $c UnM
100 1 $a Chalidabhongse, Junavit. $3 1928505
245 1 0 $a Fast motion vector estimation techniques for video coding.
300 $a 102 p.
500 $a Source: Dissertation Abstracts International, Volume: 57-07, Section: B, page: 4589.
500 $a Adviser: Chung-Chieh (Jay) Kuo.
502 $a Thesis (Ph.D.)--University of Southern California, 1996.
520 $a Motion compensated predictive coding is an effective and commonly, used method to reduce the temporal redundancy existing in image sequences. It predicts an image frame of a sequence based on the motion information between frames. Thus, motion vector (MV) estimation is a key factor in efficient video coding. Block-based motion estimation has been widely used due to its simplicity. The full search block matching algorithm (FBMA) can be performed by searching all locations in a given search area and selecting the position which gives the minimum matching error. Despite providing an optimal solution, FBMA requires a large amount of computation. The high computational complexity of MV estimation is known to be the main bottleneck in real-time encoding.
520 $a In this research, we propose a sequence of novel algorithms for fast, robust and effective block-based MV estimation by exploiting the information from spatial, spatio-temporal and multiresolution-spatio-temporal correlations of MVs. We first introduce two new algorithms using only spatial correlation to establish the basic framework. To achieve a better performance, we incorporate the temporal correlation and present two new fast algorithms based on spatio-temporal correlations. Then, we propose an ultimate fast algorithm by integrating the spatio-temporal framework with the multiresolution scheme. Its main idea is to effectively exploit the information obtained from the corresponding block at a coarser resolution level as well as the spatially and temporally adjacent blocks at the same level in order to select a good set of initial MV candidates, and then perform further local search to refine the MV result. We demonstrate with extensive experimental results that the proposed algorithm achieves a speed-up factor ranging from 150 to 310 with only 2-7% mean squared error increase and a similar rate-distortion performance in comparison with the full search algorithm.
520 $a Furthermore, the method of overlapped block motion compensation (OBMC) is reviewed and discussed. We perform experiments by applying OBMC to FBMA and our proposed fast algorithm. Comparisons are made to show that both algorithms gain approximately the same amount of performance improvement from the OBMC process.
590 $a School code: 0208.
650 4 $a Engineering, Electronics and Electrical. $3 626636
690 $a 0544
710 2 0 $a University of Southern California. $3 700129
773 0 $t Dissertation Abstracts International $g 57-07B.
790 1 0 $a Kuo, Chung-Chieh (Jay), $e advisor
790 $a 0208
791 $a Ph.D.
792 $a 1996
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=9636325