Automated Polycystic Ovarian Syndrome Identification with Follicle Recognition
J. Madhumitha, M Kalaiyarasi, S. Sakthiya Ram
Abstract
The use of ultrasound, also known as sonography, has assisted in the identification and care of infertile patients. Ultrasound imaging of the ovary's follicles provides crucial details about the ovary, such as the type of cyst, the large range of follicles, and the size of the follicles reaction to hormonal imbalance. Image Segmentation provides additional details about the region of interest in an image and accurately identifies the object and its background from the image. However, since segmentation on ultrasound images is difficult owing to noise, identification of follicles can be made simple and efficient by combining photograph preprocessing with morphological operations. SVM, KNN and Logistic Regression algorithm of Machine learning are used for classification which taking into considerations of all the specifications of PCOS Ovaries and Normal ovaries. The result obtained through the classification are compared through physical identification. The proposed hybrid method of combining all three algorithms is implemented, for which the accuracy obtained is 98%.