Wavelet Transforms and Neural Network for medical diagnosis

In this paper research work is done to introduce a novel approach of wavelet transforms and feed forward neural (FNN), to get an accurate size of the diseased tissues. Using light scattering concepts, a neural network based methodology is presented to find the size of microscopic particles. This paper tries to simplify the light scattering difficulty, as an improvement to the previous work done in this domain, a novel approach using feed forward neural networks (FNN) along with the wavelets transforms has been studied, to get an accurate size results. A comparable study has been done in order to show the size determination of particle using wavelet transforms and FNN. The size of the particle is determined from the intensity profile generated by the Mie theory concept for radius 5μm to 10 μm at an interval of 0.01 μm with varying the scattering angle from 00 to 250. The intensity vs. scattering angle profiles generated is used to deduce the sizing parameters. The accuracy of the method comes to be 93% approximately with the standard deviation calculated is 0.173 approx.