Chapter 6 Comparative Study on Mammogram Image Enhancement Methods According to the Determinant of Radiography Image Quality

摘要

Breast cancer diagnosis by analysing the mammogram becomes difficult for the low quality of X-ray image. It therefore requires a mammogram image enhancement, thus leading the extracted features to more accurate classification result. When conducting an enhancement process, it is deemed essential to consider a number of factors those are contrast sensitivity, blurring, visual noise, spotting, and detail sections that determine the quality of radiography image. This research work used 60 mammogram images from Oncology Clinic Kotabaru Yogyakarta. The images were processed by some enhancement methods based on five determinants of radiography image quality, these are smoothing by Median Filter, filtering by Butterworth Filter and Wiener Filter, denoising by Mean Filter, and histogram equalization by CLAHE. The highest PSNR was obtained by CLAHE method. Further step was to compare between Median Filter-CLAHE and Mean Filter-CLAHE. Results showed that Median Filter-CLAHE became the best enhancement method among the compared methods. This method had the lowest MSE value and the highest PSNR value. It indicated that CLAHE method able to improve the contrast of mammogram image which required high contrast sensitivity due to low contrast in chest soft tissues. The Median Filter meanwhile, was able to reduce noise and blur, thus making the mammogram image clearer. Overall, both these methods were able to enhance image quality and eligible to the determinants of radiography image quality.