Proximal Denoiser

Proximal denoisers are image processing techniques that leverage pre-trained denoising models within iterative optimization algorithms to solve inverse problems, such as image deblurring or super-resolution. Current research focuses on improving the convergence properties of these methods, particularly by exploring Plug-and-Play (PnP) algorithms and their variants, including those incorporating quasi-Newton steps or relaxed proximal gradient descent. This work aims to provide theoretical guarantees for the convergence of these algorithms while maintaining or improving the high visual quality achieved by deep learning-based denoisers, leading to more robust and efficient solutions for various image restoration tasks.