Blind Inverse Problem

Blind inverse problems aim to reconstruct an unknown signal or image from incomplete or distorted measurements when the nature of the distortion itself is also unknown. Current research heavily utilizes diffusion models, often within Expectation-Maximization (EM) frameworks or Plug-and-Play (PnP) methods, to jointly estimate both the underlying signal and the distorting operator. These approaches leverage the power of learned image denoisers and advanced architectures like transformers to achieve state-of-the-art results in applications such as image deblurring and super-resolution, MRI reconstruction, and even speaker recognition enhancement. The ability to solve blind inverse problems efficiently and accurately has significant implications for various fields requiring signal or image reconstruction from noisy or incomplete data.