CNN-Based Analysis of Ultrasound Images for PCOS Diagnosis
A. Suresh Kumar, S. Annamalai, M. Kumaresan, P. Manikandan, Ramesh Sekaran, Aditya Pai H
Abstract
The most common endocrinological condition and a major contributor to an ovulatory infertility in women worldwide is polycystic ovarian disease (PCOD/PCOS). One of the most reliable methods for diagnosing PCOS and developing an effective treatment plan for patients with this illness is the detection of numerous cysts using ovarian ultrasonography (USG) scans. A sophisticated cyst detecting technique with computer assistance may be an effective alternative to relying on labor-intensive manual identification. This research paper presents a machine learning classification method for PCOS/PCOD prediction. The most effective outcomes are attained by integrating the “MobileNet” which is a pre-trained CNN model. Compared to other ML-based strategies now in use, the accuracy of the suggested method is greatly improved while training execution time is also decreased.