A Novel Deep Learning Technique Based Automatic Detection of Polycystic Ovary Syndrome Using Ultrasound Images

One of the most common reasons women can't conceive is a hormonal issue called polycystic ovarian syndrome (PCOS). Symptoms of polycystic ovarian syndrome include being overweight, having excessive body hair, menstrual irregularities, and acne. Prompt diagnosis is key for successful management of polycystic ovarian syndrome (PCOS). However, the diagnosis is based on the Rotterdam criterion, which includes factors like high androgen hormone levels, ovulation failure, and polycystic ovaries on ultrasonography (PCOM). The current gold standard for diagnosing polycystic ovarian syndrome (PCOS) relies on ovary ultrasonography, which requires doctors and radiologists to physically count follicles and determine their volume within the ovaries. This study presents an advanced ensemble-based strategy called Adaptive-Synthetic-Sampling. It uses the cat boost algorithm for feature selection and integrates voting and averaging methods.