Unlearned Neural Networks as Image Priors for Inverse Problems

This online lecture covers Deep Image Prior, Deep Decoder, and Deep Geometric Prior for inverse problems in imaging. The course aims to teach students about utilizing unlearned neural networks as image priors for various imaging tasks. The learning outcomes include understanding the concepts of Deep Image Prior, Deep Decoder, and Deep Geometric Prior, as well as their applications in image reconstruction and enhancement. The course teaches skills such as utilizing neural networks for image processing, implementing Deep Decoder architecture, adjusting Deep Decoder parameters, and leveraging Deep Geometric Priors for surface reconstruction. The teaching method involves a lecture format with theoretical explanations and practical examples. This course is intended for students and professionals interested in artificial intelligence, image processing, computer vision, and deep learning techniques for solving inverse problems in imaging.