Evaluasi Teknik Filtering Contrast Enhancement dan Edge Sharpening untuk Pengolahan Citra Ultrasonografi Prostat

Abstract

One of the modality to detect prostate’s abnormalities is ultrasonography (USG). Indonesia still relies on 2D USG images despite its weaknesses. In this research, image processing called Computer-Aided Diagnostic (CAD) was performed on 2D USG prostate images. This research is a basic CAD research to improve the quality of 2D USG images with various combination of image procesing methods. Evaluation of the processed images are performed by not only SNR method, but also evaluation from clinicians. Twenty prostate images were used with normal and abnormal diagnosis. Image processing methods used are median filter, wiener filter, gaussian filter, CLAHE (Contrast Limited Adaptive Histogram Equalization) contrast enhancement and image sharpening. Those methods were combined so there are three combinations to be evaluated. Among 3 combinations, gaussian filter combination has the highest SNR in normal and abnormal images. But, qualitative evaluation performed by clinicians shows that median filter combination has the best visualization of prostate capsule and prostate size. Evaluation from comparing pixel value and prostate size between normal and abnornal images shows that median filter combination has the biggest differences, 13.11, makes it the best combinaton to show different visualization of normal and abnormal prostate.