Image Reconstruction Using Analysis Model Prior

摘要: The analysis model has been previously exploited as an alternative to the classical sparse synthesis for designing image reconstruction methods. Applying a suitable operator on of interest yields cosparse outcome which enables us reconstruct from undersampled data. In this work, we introduce additional prior in context and theoretically study uniqueness issues terms operators general position specific 2D finite difference operator. We establish bounds minimum measurement numbers are lower than those cases without using prior. Based idea iterative cosupport detection (ICD), develop novel effective algorithm, achieving significantly better performance. Simulation results synthetic practical magnetic resonance (MR) images also shown illustrate our theoretical claims.