Title: Classification of breast cancer images using completed local ternary pattern and support vector machine

Authors: M. Kusuma Sri; E. Gomathi

Addresses: Department of Electronics and Communication Engineering, Anurag University, Hyderabad – 500088, India ' Department of Electronics and Communication Engineering, KIT-Kalaignarkarunanidhi Institute of Technology, Kannampalayam, Coimbatore – 641402, India

Abstract: Breast cancer is the major cause of deaths in women compared to other cancers. Though early detection of breast cancer reduces cancer deaths, it is a challenging task for physicians. Local binary pattern (LBP) and local ternary pattern (LTP) techniques are widely applied in texture classification applications. Since LBP is more sensitive to noise in texture classification, it needs to be improved for achieving better results. Though LTP is more robust to noise, there are few drawbacks. Completed LBP and completed local binary count techniques achieve good accuracy for texture classification, but they inherit few drawbacks of LBP. In this paper, completed LTP operator is applied on breast cancer images for better classification accuracy than LBP and completed LBP operators, by extracting sign and magnitude components. Experimental results based on breast cancer database show that the proposed technique achieved better classification accuracy than existing similar approaches.

Keywords: breast cancer image; image segmentation; texture classification; machine learning; decision tree; logistic regression; SVM.; support vector machine.

DOI: 10.1504/IJBRA.2022.121769

International Journal of Bioinformatics Research and Applications, 2022 Vol.18 No.1/2, pp.130 - 140

Received: 04 Jul 2019
Accepted: 16 Apr 2020
Published online: 07 Apr 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article