Automatic Seed Point Selection in B-Mode Breast Ultrasound Images (original) (raw)

2018

Abstract

In order to reduce the processing time and complexity to detect the boundary of lesions in breast ultrasound (BUS) images, first step is selection of region of interest (ROI), which subsequently needs selection of seed point. Seed point is starting point that lies inside the lesion region. After selection of seed point, region growing techniques are used for segmentation of lesions or for selection of region of interest. Seed point can be selected manually, but it needs human interaction. To design a fully automatic breast ultrasound computer-aided diagnosis (CAD) system, an automatic seed point selection technique is required. In this paper, an automatic seed point detection technique is proposed. This technique is applied on 108 BUS images (57 benign and 51 malignant). Results are compared with other available methods. Quantitative experiment results show that this method could find the proper seed point for 95.3% BUS Images.

madan lal hasn't uploaded this paper.

Let madan know you want this paper to be uploaded.

Ask for this paper to be uploaded.