Extraction of Tumour in Breast MRI using Joint Thresholding and Segmentation – A Study
Seifedine Kadry, Robertas Damaševičius, David Taniar, V. Rajinikanth, Isah A. Lawal
Abstract
Breast Cancer (BC) is one of the harsh conditions, which largely affects the women group. Due to its significance, a range of procedures are available for premature detection and early treatment to save the patient. The clinical level diagnosis of BC will be done using; (i) Image supported detection and (ii) Core-Needle-Biopsy (CNB) assisted confirmation. The proposed work aim to develop a computerized scheme to detect the Breast-Tumor-Section (BTS) from the beast MRI slices. This work implements a joint thresholding and segmentation methodology to enhance and extract the BTS from the 2D MRI slices. A tri-level thresholding based on Slime-Mould-Algorithm and Shannon's-Entropy (SMA+SE) is implemented to enhance the BTS and Watershed-Segmentation (WS) is implemented to mine the BTS. After extracting the BTS, a study between the BTS and Ground-Truth image is performed and the necessary Image-Performance-Values (IPV) are computed. In this work the axial, coronal and sagittal slices of 2D breast MRI are separately examined and the attained results are presented.