Generative Model for Deblurring and Noise Reduction in Robot-Assisted Laparoscopy Images

Image Recovery research field have gained more prominence in the last years with the development of new generative models for Generative Adversarial Networks (GANs) algorithms. These techniques have been used for photograph generation, processing and recovery, being one of the most relevant classes of Deep Learning algorithms for researchers from this area. This project is focused in the implementation of a generative model automatically execute deblurring, noise reduction and super-resolution in surgical images before the medical instrument tracking during Minimally Invasive Procedures, using this algorithm to improve the execution of robots.