Abstract
In this paper certain simple procedures are presented for image classification using a class of Neural Network (NN) trained by backpropagation algorithm with minimum number of failures. The representation of the images is obtained through predominant Fourier Descriptor Method (FDM). To quicken the network response (Error vs Epoch) with minimum number of Oscillations, a slope parameter is utilized in the bipolar sigmoidal activation function in the hidden and output layers. The above scheme is applied to a Feed-Forward Neural Network (FFNN) for image classification. The performance of the network in terms of local minima, learning speed is evaluated for the choices of slope parameter, learning rate, momentum factor and compared with the traditional BP Algorithm.
Indexing terms:
Additional information
Notes on contributors
S N Sivanandam
S N Sivanandam received his PhD (1982), MSc (Engg) (1966), and BE degree in Electrical and Electronics Engineering (1964) from the Madras University, Madras. His major field of research has been in the field of Control Systems particularly in analyzing the stability of Linear and Non-linear systems using Neural Networks and Fuzzy Logic. His other research areas include image Processing, Signal Processing, System Modeling, Model Reduction and Genetic Algorithms. He is presently working as the Head of the Department of Electrical and Electronics Engineering, PSG College of Technology, Coimbatore.
M Paulraj
M Paulraj received his ME degree in Computer Science Engineering (1991), BE degree in Electrical Engineering (1983) from Madras University, Madras. His major field of specialization has been in Neural Networks, Fuzzy Logic and Image Processing. He is currently a research scholar at PSG College of Technology and working as a Lecturer at Government College of Technology, Coimbatore.