- Research Article
- Open Access
A Reconfigurable FPGA System for Parallel Independent Component Analysis
EURASIP Journal on Embedded Systemsvolume 2006, Article number: 023025 (2006)
A run-time reconfigurable field programmable gate array (FPGA) system is presented for the implementation of the parallel independent component analysis (ICA) algorithm. In this work, we investigate design challenges caused by the capacity constraints of single FPGA. Using the reconfigurability of FPGA, we show how to manipulate the FPGA-based system and execute processes for the parallel ICA (pICA) algorithm. During the implementation procedure, pICA is first partitioned into three temporally independent function blocks, each of which is synthesized by using several ICA-related reconfigurable components (RCs) that are developed for reuse and retargeting purposes. All blocks are then integrated into a design and development environment for performing tasks such as FPGA optimization, placement, and routing. With partitioning and reconfiguration, the proposed reconfigurable FPGA system overcomes the capacity constraints for the pICA implementation on embedded systems. We demonstrate the effectiveness of this implementation on real images with large throughput for dimensionality reduction in hyperspectral image (HSI) analysis.
Hyvärinen A, Oja E: A fast fixed-point algorithm for independent component analysis. Neural Computation 1997,9(7):1483-1492. 10.1162/neco.19188.8.131.523
Bartlett M, Sejnowski T: Viewpoint invariant face recognition using independent component analysis and attractor networks. In Advances in Neural Information Processing Systems 9. MIT Press, Cambridge, Mass, USA; 1997:817-823.
Lennon M, Mercier G, Mouchot MC, Hubert-Moy L: Independent component analysis as a tool for the dimensionality reduction and the representation of hyperspectral images. Proceedings of IEEE International Geoscience and Remote Sensing Symposium (IGARSS '01), July 2001, Sydney, NSW, Australia 6: 2893-2895.
Comon P: Independent component analysis, a new concept? Signal Processing 1994,36(3):287-314. special issue on High-Order Statistics 10.1016/0165-1684(94)90029-9
Lee T-W, Lewicki MS, Sejnowski TJ: ICA mixture models for unsupervised classification of non-Gaussian classes and automatic context switching in blind signal separation. IEEE Transactions on Pattern Analysis and Machine Intelligence 2000,22(10):1078-1089. 10.1109/34.879789
Cohen MH, Andreou AG: Analog CMOS integration and experimentation with an autoadaptive independent component analyzer. IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing 1995,42(2):65-77. 10.1109/82.365346
Cho K-S, Lee S-Y: Implementation of infomax ICA algorithm with analog CMOS circuits. Proceedings of the 3rd International Conference on Independent Component Analysis and Blind Signal Separation, December 2001, San Diego, Calif, USA 70-73.
Celik A, Stanacevic M, Cauwenberghs G: Mixed-signal real-time adaptive blind source separation. Proceedings of IEEE International Symposium on Circuits and Systems (ISCAS '04), May 2004, Vancouver, Canada 5: 760-763.
Cauwenberghs G: Neuromorphic autoadaptive systems and independent component analysis. In Tech. Rep. N00014-99-1-0612. Johns Hopkins University, Baltimore, Md, USA; 2003. http://bach.ece.jhu.edu/gert/yip/
Bouldin D: Developments in design reuse. University of Tennessee, Knoxville, Tenn, USA; 2001.
Lim AB, Rajapakse JC, Omondi AR: Comparative study of implementing ICNNs on FPGAs. Proceedings of International Joint Conference on Neural Networks (IJCNN '01), July 2001, Washington, DC, USA 1: 177-182.
Nordin A, Hsu C, Szu H: Design of FPGA ICA for hyperspectral imaging processing. Wavelet Applications VIII, April 2001, Orlando, Fla, USA, Proceedings of SPIE 4391: 444-454.
Sattar F, Charayaphan C: Low-cost design and implementation of an ICA-based blind source separation algorithm. Proceedings of the 15th Annual IEEE International ASIC/SOC Conference, September 2002, Rochester, NY, USA 15-19.
Wei Y, Charoensak C: FPGA implementation of non-iterative ICA for detecting motion in image sequences. Proceedings of the 7th International Conference on Control, Automation, Robotics and Vision (ICARCV '02), December 2002, Singapore 3: 1332-1336.
Yamaguchi T, Itoh K: An algebraic solution to independent component analysis. Optics Communications 2000,178(1):59-64. 10.1016/S0030-4018(00)00642-8
Cover TM, Thomas JA: Element of Information Theory. John Wiley & Sons, New York, NY, USA; 1991.
Du H, Qi H, Peterson GD: Parallel ICA and its hardware implementation in hyperspectral image analysis. Independent Component Analyses, Wavelets, Unsupervised Smart Sensors, and Neural Networks II, April 2004, Orlando, Fla, USA, Proceedings of SPIE 5439: 74-83.
Leong PHW, Leong MP, Cheung OYH, et al.: Pilchard—a reconfigurable compouting platform with memory slot interface. Proceedings of the 9th Annual IEEE Symposium on Field-Programmable Custom Computing Machines (FCCM '01), April-May 2001, Rohnert Park, Calif, USA 170-179.
Chiang S-S, Chang C-I, Ginsberg IW: Unsupervised hyperspectral image analysis using independent component analysis. Proceedings of IEEE International Geoscience and Remote Sensing Symposium (IGARSS '00), July 2000, Honolulu, Hawaii, USA 7: 3136-3138.
Landgrebe D: Some fundamentals and methods for hyperspectral image data analysis. Systems and Technologies for Clinical Diagnostics and Drug Discovery II, January 1999, San Jose, Calif, USA, Proceedings of SPIE 3603: 104-113.
Du H, Qi H, Wang X, Ramanath R, Snyder WE: Band selection using independent component analysis for hyperspectral image processing. Proceedings of the 32nd Applied Imagery Pattern Recognition Workshop (AIPR '03), October 2003, Washington, DC, USA 93-98.
NASA Jet Propulsion Laboratory, California Institute of Technology, AVIRIS concept, 2001, http://aviris.jpl.nasa.org/html/aviris.concept.html