Abstract
Markov random field models provide a robust formulation for the stereo vision problem of inferring three-dimensional scene geometry from two images taken from different viewpoints. One of the most advanced algorithms for solving the associated energy minimization problem in the formulation is belief propagation (BP). Although BP provides very accurate results in solving stereo vision problems, the high computational cost of the algorithm hinders it from real-time applications. In recent years, multicore architectures have been widely adopted in various industrial application domains. The high computing power of multicore processors provides new opportunities to implement stereo vision algorithms. This article examines and extracts the parallelisms in the BP method for stereo vision on multicore processors. This article shows that parallelism of the algorithm can be efficiently utilized on multicore processors. The results show that parallelization on multicore processors provides a speedup for the BP algorithm of almost 15 times compared to the single-processor implementation on the PPE of the Cell BE. The experimental results also indicate that a frame rate of 6.5 frames/second is possible when implementing the parallelized BP algorithm on the multicore processor of Cell BE with one PPE and six SPEs.
- Bader, D. A., Agarwal, V., and Madduri, K. 2007. On the design and analysis of irregular algorithms on the cell processor: A case study of list ranking. In Proceedings of the IEEE International Parallel and Distributed Processing Symposium.Google Scholar
- Boykov, Y., Veksler, O., and Zanih, R. 2001. Fast approximate energy minimization via graph cuts. IEEE Trans. Pattern Anal. Mach. Intell. 23, 1222--1239. Google Scholar
Digital Library
- Brunton, A., Shu, C., and Roth, G. 2006. Belief propagation on the GPU for stereo vision. In Proceedings of the 3rd Canadian Conference on Computer and Robot Vision. 76--81. Google Scholar
Digital Library
- Chandy, K. M. and Misra, J. 1988. Parallel Program Design: A Foundation. Addison Wesley Publishing Company, Inc., Reading, MA. Google Scholar
Digital Library
- Chen, T., Sura, Z., O’Brien, K. M., and O’Brien, J. K. 2006. Optimizing the use of static buffers for DMA on a Cell chip. In Proceedings of the 19th International Workshop onLanguages and Compilers for Parallel Computing (LCPC’06), Revised Papers. Lecture Notes in Computer Science, vol. 4382, 314--329. Google Scholar
Digital Library
- Chen, T., Lin, H., Zhang, T., O’Braien, K. M., and O’Brien, J. K. 2008. Parallelization of AdaBoost algorithm on multi-core processors. In Proceedings of the IEEE Workshop on Signal Processing Systems. 275--280.Google Scholar
Cross Ref
- Felzenszwalb, P. F. and Huttenlocher, D. P. 2006. Efficient belief propagation for early vision. Int. J. Comput. Vision 70, 1, 41--54. Google Scholar
Digital Library
- Gschwind, M. 2007. The Cell broadband engine: Exploiting multiple levels of parallelism in a chip multiprocessor. Int. J. Paral. Process. 35, 3, 233--262. Google Scholar
Digital Library
- Hartley, R. I. 1999. Theory and practice of projective rectification. Int. J. Comput. Vision 35, 2, 115--127. Google Scholar
Digital Library
- IBM. 2007. Cell Broadband Engine Programming Handbook version 1.1. IBM.Google Scholar
- Kang, S. and Bader, D. A. 2008. Optimizing JPEG2000 still image encoding on the cell broad-band engine. In Proceedings of the 37th International Conference on Parallel Processing. 83--90. Google Scholar
Digital Library
- Kim, Y., Park, S., Chen, C., and Jeong, H. 2006. Real-time architecture of stereo vision for robot eye. In Proceedings of 8th International Conference on Signal Processing.Google Scholar
- Kindermann, R. and Snell, J. L. 1980. Markov Random Fields and Their Applications. AMS.Google Scholar
- Lai, C.-H., Hsieh, K.-Y., Lai, S.-H., and Lee, J. K. 2008. Parallelization of belief propagation method on embedded multicore processors for stereo vision. In Proceedings of IEEE/ACM/IFIP Workshop on Embedded Systems for Real-Time Multimedia. 39--44.Google Scholar
- Linderman, M. D., Collins, J. D., Wang, H., and Meng, T. H. 2008. Merge: A programming model for heterogeneous multi-core systems. In Proceedings of the 13th International Conference on Architectural Support for Programming Languages and Operating Systems. ACM, New York, NY, 287--296. Google Scholar
Digital Library
- McCool, M. D. 2006. Data-parallel programming on the Cell BE and the GPU using the rapid-mind development platform. In Proceedings of the GSPx Multicore Applications Conference.Google Scholar
- Mendiburu, A., Santana, R., Lozano, J. A., and Bengoetxea, E. 2007. A parallel framework for loopy belief propagation. In Proceedings of the Genetic and Evolutionary Computation Conference. 2843--2850. Google Scholar
Digital Library
- Miura, J., Negishi, Y., and Shirai, Y. 2002. Mobile robot map generation by integrating omnidirectional stereo and laser range finder. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems. 250--255.Google Scholar
- Murray, D. and Little, J. J. 2000. Using real-time stereo vision for mobile robot navigation. Auton. Robots 8, 2, 161--171. Google Scholar
Digital Library
- Pearl, J. 1988. Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan Kaufmann Publishers Inc. Google Scholar
Digital Library
- Petrini, F., Fossum, G., Fernández, J., Barbanescu, A. L., Kistler, M., and Perrone, M. 2007. Multicore surprise: Lessons learned from optimizing Sweep3D on the Cell broadband engine. In Proceedings of the IEEE International Parallel and Distributed Processing Symposium. 1--10.Google Scholar
- Porta, J., Verbeek, J., and Kröse, B. 2004. Active appearance-based robot localization using stereo vision. Auton. Robots 18, 1, 59--80. Google Scholar
Digital Library
- Rofouei, M., Moazeni, M., and Sarrafzadeh, M. 2008. Fast GPU-based space-time correlation for activity recognition in video sequences. In Proceedings of the IEEE Workshop on Embedded Systmes for Real-time Multimedia. 33--38.Google Scholar
- Singh, A. K. 1991. Parallel programming: Achieving portability through abstraction. In Proceedings of the 11th International Conference on Distributed Computing Systems.Google Scholar
- Sun, J., Zheng, N.-N., and Shum, H.-Y. 2003. Stereo matching using belief propagation. IEEE Trans. Pattern Anal. Mach. Intel. 25, 7, 787--800. Google Scholar
Digital Library
- Wiess, Y. and Freeman, W. T. 2001. On the optimality of solutions of the max-product belief-propagation algorithm in arbitrary graphs. IEEE Trans. Inf. Theory 47, 2, 723--735. Google Scholar
Digital Library
- Yang, Q., Wan, L., Yang, R., Wang, S., Liao, M., and Nistér, D. 2006. Real-time global stereo matching using hierarchical belief propagation. In Proceedings of the British Machine Vision Conference.Google Scholar
- Yedidia, J. S., Freeman, W. T., and Wiess, Y. 2005. Constructing free-energy approximations and generalized belief propagation algorithms. IEEE Trans. Inf. Theory 51, 7, 2282--2312. Google Scholar
Digital Library
Index Terms
Parallelization of Belief Propagation on Cell Processors for Stereo Vision
Recommendations
Parallelization Strategies and Performance Analysis of Media Mining Applications on Multi-Core Processors
This paper studies how to parallelize the emerging media mining workloads on existing small-scale multi-core processors and future large-scale platforms. Media mining is an emerging technology to extract meaningful knowledge from large amounts of ...
Optimization and Parallelization of Emedge3D on Shared Memory Architecture
SYNASC '13: Proceedings of the 2013 15th International Symposium on Symbolic and Numeric Algorithms for Scientific ComputingThis paper presents a study of techniques used to speedup a scientific simulation code. The techniques include sequential optimizations as well as the parallelization with OpenMP. This work is carried out on two different multicore shared memory ...
Stereo Matching Using Belief Propagation
In this paper, we formulate the stereo matching problem as a Markov network and solve it using Bayesian belief propagation. The stereo Markov network consists of three coupled Markov random fields that model the following: a smooth field for depth/...






Comments