BAŘINA David, KULA Michal, MATÝŠEK Michal and ZEMČÍK Pavel. Accelerating Discrete Wavelet Transforms on GPUs. In: International Conference on Image Processing (ICIP). Beijing: IEEE Signal Processing Society, 2017, pp. 27072710. ISBN 9781509021758. 
Publication language:  english 

Original title:  Accelerating Discrete Wavelet Transforms on GPUs 

Title (cs):  Akcelerace diskrétní vlnkové transformace na GPU 

Pages:  27072710 

Proceedings:  International Conference on Image Processing (ICIP) 

Conference:  International Conference on Image Processing (ICIP) 2017 

Place:  Beijing, CN 

Year:  2017 

ISBN:  9781509021758 

DOI:  10.1109/ICIP.2017.8296774 

Publisher:  IEEE Signal Processing Society 

Files:  


Keywords 

discrete wavelet transform, image processing, synchronization, graphics processors 
Annotation 

The paper presents new nonseparable scheme for computation of the discrete wavelet transform on graphics cards. 
Abstract 

The twodimensional discrete wavelet transform has a huge number of applications in imageprocessing techniques. Until now, several papers compared the performance of such transform on graphics processing units (GPUs). However, all of them only dealt with lifting and convolution computation schemes. In this paper, we show that corresponding horizontal and vertical lifting parts of the lifting scheme can be merged into nonseparable lifting units, which halves the number of steps. We also discuss an optimization strategy leading to a reduction in the number of arithmetic operations. The schemes were assessed using the OpenCL and pixel shaders. The proposed nonseparable lifting scheme outperforms the existing schemes in many cases, irrespective of its higher complexity. 
BibTeX: 

@INPROCEEDINGS{
author = {David Ba{\v{r}}ina and Michal Kula and Michal
Mat{\'{y}}{\v{s}}ek and Pavel Zem{\v{c}}{\'{i}}k},
title = {Accelerating Discrete Wavelet Transforms on GPUs},
pages = {27072710},
booktitle = {International Conference on Image Processing (ICIP)},
year = {2017},
location = {Beijing, CN},
publisher = {IEEE Signal Processing Society},
ISBN = {9781509021758},
doi = {10.1109/ICIP.2017.8296774},
language = {english},
url = {http://www.fit.vutbr.cz/research/view_pub.php?id=11443}
} 