Compressed Sensing via Measurement-Conditional Generative Models
This work addresses compressed sensing for signal recovery, offering a simple and effective improvement over existing methods, though it is incremental as it builds on prior generator-based approaches.
The paper tackles the problem of compressed sensing by proposing a framework that allows a pre-trained generator to use measurement information for prior learning, resulting in more accurate signal predictions. Experiments show uniformly superior performance with large margins, reducing reconstruction error by up to an order of magnitude in some cases.
A pre-trained generator has been frequently adopted in compressed sensing (CS) due to its ability to effectively estimate signals with the prior of NNs. In order to further refine the NN-based prior, we propose a framework that allows the generator to utilize additional information from a given measurement for prior learning, thereby yielding more accurate prediction for signals. As our framework has a simple form, it is easily applied to existing CS methods using pre-trained generators. We demonstrate through extensive experiments that our framework exhibits uniformly superior performances by large margin and can reduce the reconstruction error up to an order of magnitude for some applications. We also explain the experimental success in theory by showing that our framework can slightly relax the stringent signal presence condition, which is required to guarantee the success of signal recovery.