Implementation of the DWT using intel IA-32 SIMD extensions

Authors:
Enrique P. Martin;Jose Salvado
Affiliations:
Electrical Engineering Department, Escola Superior de Tecnologia, Instituto Politécnico de Castelo Branco, Portugal;Electrical Engineering Department, Escola Superior de Tecnologia, Instituto Politécnico de Castelo Branco, Portugal
Venue:
MAMECTIS'08 Proceedings of the 10th WSEAS international conference on Mathematical methods, computational techniques and intelligent systems
Year:
2008

Citing 6
Cited 0

A Theory for Multiresolution Signal Decomposition: The Wavelet Representation

IEEE Transactions on Pattern Analysis and Machine Intelligence
Ten lectures on wavelets

Ten lectures on wavelets
MMX Technology Extension to the Intel Architecture

IEEE Micro
AMD 3DNow! Technology: Architecture and Implementations

IEEE Micro
Implementing Streaming SIMD Extensions on the Pentium III Processor

IEEE Micro
Floating-to-fixed-point conversion for digital signal processors

EURASIP Journal on Applied Signal Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents some results on the implementation of the DWT (DWT-1) trough the lifting scheme by using general purpose processor SIMD extensions. We perform image analysis and reconstruction up to 3 levels of decomposition, using the DWT factored into lifting steps for the 9/7 wavelet filter pair. The algorithm was implemented in "C" code and evaluated in terms of performance and image degradation. Three approaches were used: floating-point representation, integer fixed-point representation and SIMD extensions integer code. The results obtained when compared to floating-point code implementation, indicate that the processing time for fixed-point is around 54% and SIMD extensions code is around 24.2%. The average PSNR results are also better for fixed-point and SIMD extensions than with floating-point code implementation.