Classification of Ovarian Cyst Using Regularized Convolution Neural Network with Data Augmentation Techniques

PCOS-polycystic ovary syndrome is one of the prevalent hormonal disorders which has currently affected women populations around the age group of 22-45, in their reproductive cycle. It has been widely observed that PCOS leads to infertility. Diagnosis of infertile has proceeded by using ultrasound images of follicles present in the ovary and further examined by the features like the size of the follicles, number of follicles, age group of patients, and the hormonal test. Based on the features, ovaries are classified into three categories like Normal ovary, Cystic ovary, and PolyCystic ovary. Usually, the diameter of a follicle is more than 2-9 mm, and the count of the follicles is more than 12, then it is considered polycystic ovary. In this paper, the classification of the ovarian cyst is implemented by using the regularized CNN method. In additionally, the justification of the classification process also improved with the data augmentation method and more droplet layer techniques for better accuracy. In the proposed algorithm, the performance of the combined procedure is evaluated with the objective type of metrics and shows the accurate detection of the follicle and leads to conclude the classification of ovarian cyst.