SDIP: Self-reinforcement deep image prior framework for image processing

Abstract

Deep image prior (DIP) proposed in recent research has revealed the inherent trait of convolutional neural networks (CNN) for capturing substantial low-level image statistics priors. This framework efficiently addresses the inverse problems in image processing and has induced extensive applications in various domains. In this paper, we propose the self-reinforcement deep image prior (SDIP) as an improved version of the original DIP. We observed that the changes in the DIP networks’ input and output are highly correlated during each iteration. SDIP efficiently utilizes this discovery in a reinforcement learning manner, where the current iteration’s output is utilized by a steering algorithm to update the network input for the next iteration, guiding the algorithm towards improved results. Experimental results across multiple applications demonstrate that our proposed SDIP framework offers improvement compared to the original DIP method, especially when the corresponding inverse problem is highly ill-posed.