Local Binary Pattern Based Texture Analysis in Real Time using a Graphics Processing Unit (long version)

This paper presents a novel implementation of the “Local Binary Pattern” (LBP) texture analysis operator on a consumer-grade graphical processing unit (GPU), which yields a 14 to 18-fold run time reduction compared to standard CPU implementations. The obtained result allows the application of this powerful texture analysis method for tasks requiring both high resolution and real time performance, such as mobile outdoor robotics or industrial inspection, without the need for specialized and thus costly hardware components.

[1]  Matti Pietikäinen,et al.  Robust Texture Classification by Subsets of Local Binary Patterns , 2000, ICPR.

[2]  Christopher Rasmussen,et al.  Combining laser range, color, and texture cues for autonomous road following , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[3]  Karl Iagnemma,et al.  Terrain Classification and Classifier Fusion for Planetary Exploration Rovers , 2008, 2007 IEEE Aerospace Conference.

[4]  Matti Pietikaeinen,et al.  Texture-based paper characterization using nonsupervised clustering , 2003, International Conference on Quality Control by Artificial Vision.

[5]  Topi Mäenpää,et al.  The local binary pattern approach to texture analysis - extensions and applications , 2003 .

[6]  A. Paasio,et al.  Dedicated hardware for parallel extraction of local binary pattern feature vectors , 2005, 2005 9th International Workshop on Cellular Neural Networks and Their Applications.

[7]  Matti Pietikäinen,et al.  A comparative study of texture measures with classification based on featured distributions , 1996, Pattern Recognit..

[8]  Ibrahim Halatci.  Vision-based terrain classification and classifier fusion for planetary exploration rovers , 2006 .

[9]  M. Topi,et al.  Robust texture classification by subsets of local binary patterns , 2000, Proceedings 15th International Conference on Pattern Recognition. ICPR-2000.