PCOS (POLYCYSTIC OVARIAN SYNDROME) DETECTION USING DEEP LEARNING









Abstract

Predicting Polycystic ovarian syndrome (PCOS) is a combination of symptoms caused by high levels of androgens in women. PCOS is caused by a mix of hereditary and environmental factors, which are common disorders such as atherosclerosis, hirsutism, acne, and hyperandrogenism, as well as recurrent infertility. According to recent studies, almost 18 percent of Indian women are affected by this disease. Doctors manually inspected ultrasound scans to establish which ovary was affected, but they couldn't tell if the cyst was benign, PCOS-related, or cancerous. DCNN-based algorithms are proposed in this study, and code for PCOS categorization is created using Python programming, and they are filled with blood or fluid using ultrasound images. The work employs DCNN-based image classification to classify PCOS in the dataset.That is, the study is based on a dataset of PCOS-related illnesses that has been trained. Finally, using performance settings, the test dataset is used to execute feature extraction and measure accuracy. PCOS (Polycystic Ovary Syndrome) is a hormonal disorder that affects many women in their reproductive years and has been associated to infertility, diabetes, and cardiovascular disease. To diagnose the condition, the bulk of imaging parameters are used. Ultrasound imaging has become an important tool for diagnosing PCOS. The normal appearance of the image becomes increasingly difficult because to overlapping follicles, inherent noise of the equipment, and a lack of operator comprehension because it is primarily an experience-based operation, making the diagnosis procedure time demanding. As a result of the aforementioned situations, cyst detection accuracy is affected. Early and accurate diagnosis of anomalies in the female reproductive system is critical prior to the treatment process to avoid infertility. This paper covers several methodologies proposed so far in terms of reducing speckle noise, extracting region of interest using segmentation, and picture classification in order to achieve maximum accuracy in cyst diagnosis in a short period of time.


Modules


Algorithms


Software And Hardware