Multi-ROI segmentation for effective texture features of mammogram images

Abstract

Image segmentation is an emerging requirement in digital mammography for segmenting the mammogram images. It is performed by region of interest (ROI), which is useful for the extraction of suspected tumors or masses from the mammogram images. In the classification of mammogram images, the number of training images are required; in such cases, ROI demands more computational time. This paper is focused on this problem for resolving the time complexity issues with the proposed multi-ROI technique. Experiment is conducted on benchmarked datasets for demonstrating the efficiency of proposed multi-ROI with the comparison of the existing segmentation method.

Subject Classification:

• 62H35
• 62M40

Keywords:

• Digital mammography
• Calcifications
• Enhancement techniques
• Pre-processing