GPU-Accelerated Texture Analysis Using Steerable Riesz Wavelets

• Published in: 2016 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP) pp. 431 - 434
• Main Authors: Vizitiu, Anamaria, Itu, Lucian Mihai, Joyseeree, Ranveer, Depeursinge, Adrien, Muller, Henning, Suciu, Constantin
• Format: Conference Proceeding Journal Article
• Language: English
• Published: IEEE 01.02.2016
• Subjects: Computing time; GPU; Graphics processing units; Image processing; Instruction sets; Kernel; Mathematical models; Medical imaging; Memory management; Networks; parallel processing; Registers; Run time (computers); Support vector machines; Surface layer; Texture; texture analysis; Training; Wavelet analysis
• ISSN: 2377-5750
• DOI: 10.1109/PDP.2016.105

Visual pattern recognition is a key research topic in the field of image processing and computer vision. Texture analysis based on steerable Riesz wavelets is powerful, but requires computing pixel-wise operations resulting in a run time in the order of days when large volumes of data are processed. To overcome this limitation we propose a Graphics Processing Unit (GPU) based solution. A standard CPU version is used as starting point for the development of baseline GPU versions. To further increase the performance, and to overcome compute and memory limitations we apply a series of optimization techniques, leading to five versions in total. The best performing GPU solution ensures a speed-up of 93× for the parallelized section of the application and of 29.6× for the entire application. Furthermore, we show that a higher Riesz order and/or a higher image resolution further increases the speed-up.