APPLICATION OF TYPE-2 FUZZY LOGIC TO REMOVE NOISE FROM COLOUR IMAGES BY USING BACTERIAL FORAGING

Abstract

Usually during acquisition or transmission of an image, it gets contaminated with the impulse noise. In this paper, we present a novel application of type-2 fuzzy logic to the design of an image processing operator called an impulse detector to remove this impulse noise from images. The type-2 fuzzy logic based impulse detector can be used to guide impulse noise removal filters to significantly improve their filtering performance and enhance their final images. The structure of the proposed impulse detector is based on two 3-input and 1-output first order Sugeno type, interval type-2 fuzzy inference system. The values of the internal parameters of the type-2 fuzzy membership functions of the systems are determined by experiment. When a noisy image is passed through the detector, its job is to determine which of the pixels are noisy and which are not and then filter is applied on these noisy pixels to output a restored image. Two advantages are addressed through the use of detector. One is, not all the pixels of an image are noisy. So, it helps in recognizing the noisy pixels. Other is, filter need not to be applied on complete image but is applied only on noisy pixels. Application of filter on complete image degrades the quality of image. So, the usage of detector before filtering helps in restoring the quality of image as well as it cuts cost. The performance of the impulse detector is evaluated by using it in combination with median impulse noise filter on four different popular test images under realistic noise condition of 30%. The results demonstrate that the type-2fuzzy logic based impulse detector can be used as an efficient tool to effectively improve the performances of impulse noise filters and reduce the impulse noise undesirable distortion effects. But it has been shown that better performance can be expected if the internal parameters of the fuzzy inference system are optimized. The internal parameters of the type-2 fuzzy membership functions of the systems are determined by training. Bacterial Foraging Optimization Technique has been used for the training purpose. In order to determine the ideal behaviour of the impulse detector, an ideal detector has been developed which outputs the 100% accurate results which can be used by the optimization algorithm for the computation purpose. Also, it helps in comparing the results by the detector with and without optimization.