Contrast Enhancement of Mammograms Images Based on Hybrid Processing

Abstract

This paper introduces a new enhancement algorithm based on combination of different processing techniques. The method uses different methods at different stages of processing. In the beginning input image given to the algorithm is a portable gray map image and then Gaussian low pass filter is used to decompose the input image into low and high frequency components. On low frequency components we apply mathematical morphological operations and on high frequency components we apply edge enhanced algorithm. After this we combine processed low and high frequency components to get an enhanced image. Enhanced image is having better contrast and edge visibility comparing to the original image, but it contains noises. Wavelet transform is used to denoise the noisy image. The denoised image is then processed by using contrast limited adaptive histogram equalization(CLAHE) to have better edge preservation index (EPI) and contrast improvement index (EPI). The resulting image is then smoothed by passing the output image through a guided image filter(GIF).The edge preserve capacity and preservation of the naturalness of the GIF allows us to get better results.

The efficiency of any service or product, especially those related to medical field depends upon its applicability. The applicability for any service or products can be achieved by applying the basic principles of Software Engineering. Applicability of enhanced algorithms depends on parameters like the peak signal to noise ratio(PSNR),edge preservation index(EPI), etc. This paper focuses on introducing a model which highlights on a prototyping approach for highlighting the necessary details that will aid radiologist for the earlier detection of breast cancer.This paper also presents Design of the model, Implementation of the model and finally the results are analyzed by considering the quality metrics values like PSNR, EPI, CII.