%0 Conference Proceedings
%@isbn 978-85-7669-272-0
%T Faster methods for computing slant haar transforms of digital images
%D 1994
%A Anguh, Maurence M.,
%A Martin, Ralph R.,
%@affiliation Department of Electrical Engineering of Centro Tecnológia of Federal University of Maranhão (UFMA)
%@affiliation Department of Computing Mathematics (UWCC)
%E Freitas, Carla dal Sasso,
%E Geus, Klaus de,
%E Scheer, Sérgio,
%B Simpósio Brasileiro de Computação Gráfica e Processamento de Imagens, 7 (SIBGRAPI)
%C Curitiba
%8 9 - 11 nov. 1994
%I Sociedade Brasileira de Computação
%J Porto Alegre
%V 1
%P 125-132
%S Anais
%K Truncation Slant Haar Transform, Slant Haar Transform, Fast Slant Haar Transform, digital images, faster methods for computing, imaging.
%X A phenomenon characteristic of digital images is the presence of approximately constant or uniformly changing gray levels over a considerable distance or area. A novel method, the Truncation Slant Haar Transform (TSHT) method for computing the Slant Haar Transform (SHT) of digital images is presented which exploits this phenomenon to advantage. The TSHT method utilizes a hierarchical tree to segment and aggregate uniform image data and a matrix factorization to eliminate transform matrix redundancies. This simultaneous exploitation of inter-pixel relationships and the elimination of transform matrix redundancies produces an efficient method for computing the SHT of digital images. In one dimension with an array of N=2n data values, the TSHT method takes time between O (N) and (N log2 N), thus degenerating to the performance of the standard Fast Slant Haar Transform (FSHT) method. In two dimensions with an array of size N x N, the performance of the TSHT method is between O (N2) and O (N2 log2 N), again degenerating to that of the FSHT method in its worst case. Since coherence is a fundamental characteristic of digital images, the TSHT method is therefore superior to the FSHT method when used to compute the SHT of coherent digital images. Experimental results are presented to justify this assertion.
%9 Processamento de Imagens II
%@language en
%3 17 Faster methods for computing slant haar transforms of digital images.pdf