An Intelligent Machine Learning Approach for Ovarian Detection and Classification System using Ultrasonogram Images

Kiruthika V1

Sathiya S2

M. M. Ramya3

K. Sakthidasan Sankaran4,Email

1School of Electronics Engineering, Vellore Institute of Technology, Chennai-600127, India.   
2Department of Obstetrics and Gynaecology, Chettinad Hospital and Research Institute, Chennai-603103, India. 
3Agurchand Manmull Jain College, Chennai-600114, India.
4Deparmtent of Electronics and Communication Engineering, Hindustan Institute of Technology and Science, Chennai-603103, India 

 

Abstract

Ovulatory disorders contribute significantly for female infertility. Intelligent ovarian detection and classification is required for diagnosing and treating ovulatory disorders.  Normal and cystic images reveal similar image characteristics which makes it challenging to classify among the two. In this paper, a suitable machine learning algorithm is identified for intelligent ovarian classification. Input to classifiers is a combination of image characteristics: intensity, texture features, demographic and diagnostic features. Diagnostic and demographic features are used with a view to develop a diagnostic algorithm with better accuracy. Standard machine learning techniques like Artificial Neural Network, Support vector machine and Linear Discriminant Analysis have been investigated to analyze its performance in ovarian classification. Support vector machine was identified as the best classifier in achieving the ovarian classification with an overall average accuracy of 98%, sensitivity of 96% and specificity of 95.66%. The region of convergence also depicted a better classification of ovaries with the support vector machine classifier. The developed intelligent ovarian classification system with high accuracy will serve as a preliminary investigating technique in diagnosis of infertility and will support clinicians in decision making.

An Intelligent Machine Learning Approach for Ovarian Detection and Classification System using Ultrasonogram Images