Employing Histogram Equalization Enhancement Technique to Segment Different Medical Images of Three Organs

Abstract

Histogram equalization is a pixel based enhancement technique in spatial domain. In this work, this technique employed to achieve segmentation process for different modalities of medical images: mammography of breast, MRI and CT scan images of brain and liver in order to detect tumors and other abnormalities like Haemorrhage in MRI brain images. In addition, a hybrid method based on clustering and histogram equalization was proposed to improve the performance of K-means clustering and histogram equalization for isolating and extracting tumors in the adopted experimental images. The results showed that the proposed histogram equalization method, succeeded adequately to segment, detect and extract tumors and abnormality regions, so it could be used as segmentation technique beside its function in images enhancement to improve their appearance. As well as, the hybrid technique succeeded to isolate and extract tumors and abnormality regions in the tested medical images according to the consultation of the radiologist. The results of the proposed methods were encouraging and in good agreement with the radiologist delineation. The percent relative difference of the calculated surface area it was found ranging from (0.01508 to 0.32281)%.