Compressed Sensing System with Simultaneous Optimization of Sensing Matrix and Sparsifying Dictionary Applied in Gastrointestinal Images for Remote Diagnosis

This paper copes with the compressed sensing (CS) technique which can be applied in remote diagnose on gastrointestinal disease. The images from gastrointestinal endoscopes can be compressed by a designed sensing matrix so that the transmission rate can be improved. Then the estimate images can be recovered in the designed sensing matrix and trained dictionary by using proper recovery algorithms. The algorithms are designed for optimizing sensing matrix and sparsifying dictionary alternatively, which will bring accurate recovery performance. The simulations demonstrate the superiority in the application of remote diagnose by adopting CS technique.