國立高雄大學圖資館 |

Language: English

Back

Best basis algorithms for tree-struc...

Ho, Hsin-Han.

Best basis algorithms for tree-structured Haar transforms and their application.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Best basis algorithms for tree-structured Haar transforms and their application.
Author:	Ho, Hsin-Han.
Description:	98 p.
Notes:	Source: Dissertation Abstracts International, Volume: 71-05, Section: B, page: 3225.
Notes:	Adviser: Sanjit K. Mitra.
Contained By:	Dissertation Abstracts International71-05B.
Subject:	Engineering, Computer.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3398795
ISBN:	9781109726756

Best basis algorithms for tree-structured Haar transforms and their application.
Ho, Hsin-Han.

Best basis algorithms for tree-structured Haar transforms and their application. - 98 p.

Source: Dissertation Abstracts International, Volume: 71-05, Section: B, page: 3225.

Thesis (Ph.D.)--University of California, Santa Barbara, 2010.

In this dissertation, we report our contributions on one-dimensional tree-structured Haar (TSH) transforms, with a focus on the development of a best basis search given an information cost function for the approximation of a discrete one-dimensional signal. We begin by going from single binary tree to multiple disjoint binary trees, which we call multiple TSH transforms. Next, we use a lattice structure to visualize hierarchical dependencies among the dictionary of discrete TSH functions in multiple TSH transforms. We also outline a new, space-efficient implementation of fast transform computation based on generating a parameter table fully specifying the butterfly network. We then provide best basis algorithms for TSH transforms and multiple TSH transforms, respectively. Both algorithms are later combined with greedy approximation to form a heuristic solution for B-term approximation under normalized &ell;p-norm error metric and normalized &ell; p-norm relative error metric with sanity bound. Lastly, we mention one potential application of our heuristic solution and provide performance comparisons with state-of-the-art algorithms.

ISBN: 9781109726756Subjects--Topical Terms:

384375
Engineering, Computer.

Best basis algorithms for tree-structured Haar transforms and their application.
LDR:02250nmm 2200313 4500 001 280812
005 20110119095000.5
008 110301s2010 ||||||||||||||||| ||eng d
020 $a 9781109726756
035 $a (UMI)AAI3398795
035 $a AAI3398795
040 $a UMI $c UMI
100 1 $a Ho, Hsin-Han. $3 492938
245 1 0 $a Best basis algorithms for tree-structured Haar transforms and their application.
300 $a 98 p.
500 $a Source: Dissertation Abstracts International, Volume: 71-05, Section: B, page: 3225.
500 $a Adviser: Sanjit K. Mitra.
502 $a Thesis (Ph.D.)--University of California, Santa Barbara, 2010.
520 $a In this dissertation, we report our contributions on one-dimensional tree-structured Haar (TSH) transforms, with a focus on the development of a best basis search given an information cost function for the approximation of a discrete one-dimensional signal. We begin by going from single binary tree to multiple disjoint binary trees, which we call multiple TSH transforms. Next, we use a lattice structure to visualize hierarchical dependencies among the dictionary of discrete TSH functions in multiple TSH transforms. We also outline a new, space-efficient implementation of fast transform computation based on generating a parameter table fully specifying the butterfly network. We then provide best basis algorithms for TSH transforms and multiple TSH transforms, respectively. Both algorithms are later combined with greedy approximation to form a heuristic solution for B-term approximation under normalized &ell;p-norm error metric and normalized &ell; p-norm relative error metric with sanity bound. Lastly, we mention one potential application of our heuristic solution and provide performance comparisons with state-of-the-art algorithms.
590 $a School code: 0035.
650 4 $a Engineering, Computer. $3 384375
690 $a 0464
710 2 $a University of California, Santa Barbara. $b Electrical & Computer Engineering. $3 492939
773 0 $t Dissertation Abstracts International $g 71-05B.
790 1 0 $a Mitra, Sanjit K., $e advisor
790 1 0 $a Chandrasekaran, Shivkumar $e committee member
790 1 0 $a Egiazarian, Karen O. $e committee member
790 1 0 $a Gibson, Jerry D. $e committee member
790 1 0 $a Kuo, C.-C. J. $e committee member
790 $a 0035
791 $a Ph.D.
792 $a 2010
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3398795