Digital Hardware Pulse-Mode RBFNN with Hybrid On-chip Learning Algorithm Based Edge Detection.
About this article
Abstract
A hardware implementation of pulse mode Radial Basis Function Neural Network (RBFNN) with on-chip learning ability is proposed in this paper. Pulse mode presents an emerging technology in digital implementation of neural networks thanks to its higher density of integration. However, hardware on-chip learning is a difficult issue, since the back-propagation algorithm is the most used, which requires a large number of logic gates in the hardware. To overcome this problem, we apply a hybrid process, which is split into two stages. In the first one, the K-means algorithm is used to update the centers of gaussian activation functions. Thereafter, the connection weights are adjusted using the back-propagation algorithm. Details of important aspects concerning the hardware implementation are given. As illustration of the efficiency and scalability of the proposed design, we consider edge detection operation which is a very important step in image processing. In the learning step, the RBFNN was taught the Canny operator behavior. Experiential results show good approximation features. The proposed design was implemented on a virtex II PRO FPGA platform and synthesis results showed higher performances when benchmarked against conventional techniques and neural ones.
References
[1] J.K. Sing, D.K. Basu, N. Mita, and M.K. Nasipuri, “Face recognition using point symmetry distance-based rbf network,” Applied Soft Computing, vol.7,(2005), pp58-70.
[2] W. Guo, T. Qiu, H. Tang, and W. Zhang, “Performance of RBF neural networks for array processing in impulsive noise environment,” Digital Signal Processing, vol. 8, (2008),pp168-178.
[3] D. Du, K. Li, M. Fei, “A fast multi-output rbf neural network construction method,” Neurocomputing, (2010).
[4] S. Song, Z. Yu, X. Chen, “A novel radial basis function Neural network for approximation,” International Journal of Information Technology, vol.11, no.9, (2005).
[5] O. Polat, T. Yildirim, “FPGA implementation of a General Regression Neural Network: An embedded pattern classification system,” Digital Signal Processing, vol. 20, (2010),pp.881–886.
View more references (12)
[6] D.E. Van Den Bout, T.K. Miller III “A Digital Architecture Employing Stochasticism for the Simulation of Hopfield Neural Nets,” IEEE Trans. Circuits Systems, vol. 36, n.5(1989).
[7] M. Krid, A. Dammak, and D.S. Masmoudi, “Hardware implementation of pulse mode RBF neural network-based image denoising,” Int. J. Electron. Commun. (AEU), vol. 63, (2009),pp.810–820.
[8] H. Hikawa, “Frequency-Based Multilayer Neural Network with On-Chip Learning and Enhanced Neuron Characterisitcs,” IEEE Trans. Neural Networks, vol. 10, no. 3, (2003).
[9] Y. Maeda, and T. Tada, “FPGA Implementation of a Pulse Density Neural Network With Learning Ability Using Simultaneous Perturbation,” IEEE Trans. Neural Networks, vol.14, no. 3, (2003).
[10] Kaliraj & Baskar, “An efficient approach for the removal of impulse noise from the corrupted image using neural network based impulse detector,” Image and Vision Computing, vol. 28, (2003), pp.458–466.
[11] D. Casasent, and X. Chen, “New training strategies for RBF neural networks for X-ray agricultural product inspection,” Pattern Recognition, vol. 36,( 2003), pp.535- 547.
[12] M. Krid, A. Gargouri, and D. S. Masmoudi, “Hardware implementation of pulse mode RBF neural network-based image denoising,” Int. J. Innovative Computing and Applications, vol. 3, no. 4, (2011).
[13] Y. Zhao, F. E.H. Tay, F. Siong Chau, and G. Zhou, “Linearization of the scanning field for 2D torsional micromirror by RBF neural network,” Sensors and Actuators, vol. 121, pp.230–236, (2005).
[14] A. Basturk, and E. Gunay, “Efficient edge detection in digital images using a cellular neural network optimized by differential evolution algorithm,” Expert Systems with Applications, vol. 36, (2009),pp.2645–2650.
[15] M.E. Yuksel, “Edge detection in noisy images by neuro-fuzzy processing,” Int. J. Electron. Commun. (AEU), vol. 61, (2007),pp.8 –89.
[16] L.M. Reyneri, “Edge A performance analysis of pulse stream neural and fuzzy computing system,” IEEE Trans Circuits Syst, vol. 42,( 1995) pp.624–60.
[17] J. Z. Wang, z. James, "wang research group: Databases for research comparison,” [Online]. Available: http:// wang.ist.psu.edu/docs/ related. shtml