Image De-noising Method Using Median Type Filter, Fuzzy Logic and Genetic Algorithm

Abstract

The goal of de-noising the medical images is to get rid of the distortions occurred in the noisy medical images. A new methodology is proposed to overcome the impulse noise affected mammogram images by using hybrid filter (HF), fuzzy logic (FL) and genetic algorithm (GA). The above said method is implemented in three steps: The primary step includes denoising of noisy mammogram images using median filter and adaptive fuzzy median filter respectively. The intermediate step intends to compute the difference vector using the above two filters and it is then given to a fuzzy logic-based system. The system utilizes triangular membership function to generate the fuzzy rules from the computed difference vector value. The last step makes use of genetic algorithm to select the optimal rule. Peak signal to noise ratio (PSNR) value is needed to be found for each population. For obtaining the best fitness value, the new population is formed repeatedly with the help of genetic operator. The performance of the method is measured by calculating the PSNR value. The proposed implementation is tested over mammogram medical images taken from Mammogram Image Analysis Society (MIAS) database. The experimental results are compared with different exiting methods.