An Intelligent Ovarian Ultrasound Image Generation Algorithm based on Generative Adversarial Networks

A deep learning framework based on generative adversarial networks (GAN) for ovarian ultrasound (US) images synthesis is investigated. This method offers an effective solution for addressing the issue of insufficient and unbalanced data in ovarian disease research. The proposed network, built on the Triple-GAN model, can synthesize a large number of medical ovarian US images which are difficult to distinguish from the real ones. This approach effectively increases the available volume of the ultrasound images of ovarian diseases and facilitates deep learning applications in ovarian ultrasound images. It provides reliable training data replacements. The generated data were validated through professional appraisal, classification method and image metrics. The results demonstrated high credibility and quality. The feasibility of using the generated data in the intelligent classification algorithm of ovarian diseases is verified, which has important practical significance for the research and development of artificial intelligence diagnosis algorithms for various diseases in the future.