Classification of PCOS in ultrasound images using deep learning methods

Polycystic Ovary Syndrome (PCOS) stands out as a widespread health condition that notably impacts women of reproductive age. Diagnosing PCOS can be challenging because not everyone with ovarian cysts or polycystic ovaries (PCO) has PCOS; therefore, a pelvic ultrasound alone is insufficient to diagnose PCOS. The major components of the comprehensive diagnostic approach include a pelvic ultrasound in addition to blood testing for particular criteria that show PCOS. Given that PCOS is a widespread hormonal illness that is difficult to identify, using a computer to analyze blood tests, symptoms, and other criteria can provide a new, simple diagnostic technique. This research used KNN and SVM classifiers to implement a variety of machine methods. Approximately 70% of these instances go undetected on average. Previously, ultrasound pictures were utilized to segment follicles using an adaptive existing technique; however, this technique did not focus on classifying aberrant and normal follicles. For the purpose of classifying follicles as abnormal (Chocolate cysts) or normal (Clear cysts), the KNN and SVM algorithms were employed. After that, the feature extraction model and follicle segmentation are processed using this technique. Finally, KNN and SVM will be trained using these characteristics for classification. Experimental results show that the suggested model is effective in terms of accuracy, recall, and precision.