We developed an Artificial Neural Network on a Graphical Processing Unit inside a video card we used NVIDIA CUDA for this implementation. In order to design a video card based classifier we initially developed a digital circuit to recognize few characters and then we converted that circuit into its equivalent ANN classifier . Here we observed that ANN calculation has reduce to a greater extend thus speeding up the calculations and making it suitable for real-time applications. GPUs have hundreds of processing units and have a highly parallel architecture, that clearly maps to ANN since ANN is also a massively parallel system. Moreover the implementation of an Artificial Neural Network on a GPU provides improved performance as compared to CPU implementation. the GPU-based ANN is much more cost effective as compared to FPGA or ASWIC based solutions. This research aims at implementation of an Artificial Neural Network on a GPU in order to improve the performance as compared to CPU implementation.

1. R. Tom, Halfhill, "Parallel processing with CUDA", Microprocessor report. January 2008
2. Z. Luo, H. Liu and X. Wu, "Artificial Neural Network Computation on Graphic Process Unit", Neural Networks, 2005. IJCNN '05. Proceedings. 2005 IEEE International Joint Conference, vol. 1, pp. 622 – 626, 31 July-4 Aug. 2005
3. R. D. Prabhu, "GNeuron: Parallel Neural Networks with GPU", Posters, International Conference on High Performance Computing (HiPC) 2007, December 2006
4. Shuai Che , Michael Boyer, Jiayuan Meng, David Tarjan , Jeremy W. Sheaffer, Kevin Skadron, "A Performance Study of General-Purpose Applications on Graphics Processors Using CUDA", Journal of Parallel and Distributed Computing, ACM Portal, Volume 68 ,Issue 10, pp. 1370-1380 October 2008
5. Elizabeth, Thomas , "GPU GEMS: Chapter 35 Fast Virus Signature Matching on the GPU", Addison Wesley
6. Pat Hanrahan , "Why are Graphics Systems so Fast? " , PACT Keynote, Pervasive Parallelism Laboratory, Stanford University, September 14, 2009
7. Kayvon Fathalian and Mike Houston, "A closer look at GPUs", Communications of the ACM, Vol . 51 no. 10, October 2008
8. Trendy R. Hagen, Jon M. Hjelmervik, Tor Dokken, "The GPU as a high performance computational resource", Proceedings of the 21st spring conference on Computer graphics, pp. 21 – 26, 2005
9. R. H. Luke, D. T. Anderson, J. M. Keller S. Coupland, "Fuzzy Logic-Based Image Processing Using Graphics Processor Units", IFSA-EUSFLAT 2009
10. Victor Podlozhnyuk, Mark Harris , "NVIDIA CUDA Implementation of Monte Carlo Option Pricing", Tutorial in NVIDIA CUDA SDK 2. 3,
11. "CUDA Programming Guide 3. 0", Published by NVIDIA Corporations.
12. David B. Kirk and Wen-mei W. Hwu, Programming Massively Parallel Processors: A Hands-on Approach by Morgan Kaufmann (February 5, 2010),ISBN 0123814723
13. K. Oh, "GPU implementation of neural networks" Pattern Recognition, pp. 1311-1314. Vol. 37, No. 6. (June 2004).
14. Simon Haykin, "Neural Networks: A comprehensive foundation", 2nd Edition, Prentice Hall, 1998
15. Araokar, Shashank, "Visual Character Recognition using Artificial Neural Networks" Cog Prints, May 2005
16. Earl Gose ,Steve Jost, Richard Johosonbaugh, "Pattern Recognition and Image Analysis" , Eastern Economy Edition, Prentice Hall ,1999.

Gpu Nvidia Cuda Ann Classifier training Pattern Recognition.