GANvCan: “Improving cancer detection in X-rays with the help of generative models”, owner National University of Science and Technology Politehnica Bucharest (led by AI Multimedia Lab), funded by National University of Science and Technology Politehnica Bucharest, GNAC2023 ARUT, STARTING GRANT-SG (budget ~10k Eur, November 2023 - December 2024);
Objectives
Medical imaging techniques are a direct beneficiary of the development of deep learning, which is confirmed by the existence of algorithms for classifying and detecting nodules in order to detect lung cancer. The current project proposes to improve cancer nodule detection algorithms by augmenting training databases with synthetic examples, generated with algorithms such as GAN and VAE. The generated nodules are inserted and integrated into real images, increasing the number of positive examples on which detection algorithms can train, leading to better final results.
