A Fast Recovery Method of 2D Geometric Compressed Sensing Signal

摘要

This paper presents a compression method based on compressed sensing for 2D contour models. And low rank random matrixes are used to sample 2D contour models, since the models can be sparsely represented under their Laplace operators. In the recovery process, a new function is designed as the optimal objective function to replace signal’s 1-norm, while a new search direction is constructed to find the solution, and to ensure that the solution speed is equal to the speed of the linear optimization. Finally, the experimental results show that the above method, boasting advanced compression ratio and good recovery effect, is well suited for processing large data.