IV3BoostPCOS: A Comprehensive Polycystic Ovary Syndrome (PCOS) Detection Framework

Polycystic ovary syndrome (PCOS) is a prevalent gynecological illness characterized by hormonal irregularities and ovarian cyst production that affects women all over the world. Detecting numerous cysts using ovarian ultrasonography (USG) scans is a solid method for accurately diagnosing PCOS and developing treatment plans for individuals. A sophisticated computer-assisted cyst identification system is a feasible alternative to manual identification, which is pronetoe rrors. Hence, in this research, a sophisticated ultrasound image categorization method is proposed that aims at optimizing three powerful convolutional Neural Network (CNN) models to extract features from the ovarian ultrasonography (USG) images. The reduced feature set is then used to categorize the ultrasound images using the Extreme Gradient Boosting classifier (XGBoost) to categorize between PCOS and non-PCOS ovaries. The proposed methodology, which used the “Inception V3” model without top classification 1 ayers f or feature extraction a nd t he XGBoost model as an image classifier, o btained 9 9.31% a ccuracy and reduced training execution time when compared to current PCOS detection methodologies.