Breast Cancer Detection and Classification using NSST-based Modified FPCM with ANN

  • Shaik Shabbir Hussain, Ongole Gandhi, Sajida Sultana. Sk

Abstract

Breast cancer is one of the leading diseases for women in the world. It is ranked second among all types of cancers, to cause death in women. So, the early detection is necessary to reduce the mortality rate. The early detection of breast cancer and treatment leads to an increase in the survival rate of women. Mammography is a standard radiological screening technique, which is used for checking of breast cancer in women when there are no symptoms. Few machine learning (ML) approaches like support vector machine (SVM) utilized for the detection and classification of breast cancer in the literature. However, SVM produced the results with inaccurate classification. Thus, to enhance the accuracy of cancer detection and classification, this article proposed nonsubsampled shearlet transform (NSST) based modified fuzzy probabilistic C-means (MFPCM) clustering approach for detection of breast cancer. In addition, spatial gray level dependence matrix (SGLD) is utilized to extract the features from the segmented breast cancer image and then artificial neural network (ANN) which is a concept of artificial intelligence is employed for classification of type of cancer. The proposed model is trained on about 425 images (150 of benign, 150 of malignant and 125 of normal breast cancer images respectively), and obtained classification accuracy of 95.91%, sensitivity and specificity of 96.34% and 95.81%. These extensive simulation results demonstrate that proposed hybrid approach achieves enhanced classification accuracy with comparison to the existing ML-based approaches in terms of medical statistical parameters like specificity, sensitivity and F1-score as well.

Published
2020-04-27
How to Cite
Shaik Shabbir Hussain, Ongole Gandhi, Sajida Sultana. Sk. (2020). Breast Cancer Detection and Classification using NSST-based Modified FPCM with ANN. International Journal of Advanced Science and Technology, 29(8s), 1567 - 1576. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/12569