SDIP: Self-Reinforcement Deep Image Prior Framework for Image Processing
This work addresses stability issues in image processing for researchers and practitioners, but it is incremental as it builds directly on the existing DIP framework.
The paper tackles the instability of the Deep Image Prior (DIP) method in image processing by proposing a self-reinforcement framework (SDIP) that uses reinforcement learning to guide network updates, resulting in improved performance across multiple applications compared to DIP and other state-of-the-art methods.
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. However, as the whole algorithm is initialized randomly, the DIP algorithm often lacks stability. Thus, this method still has space for further improvement. 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 trait 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 toward improved results. Experimental results across multiple applications demonstrate that our proposed SDIP framework offers improvement compared to the original DIP method and other state-of-the-art methods.