An improved gradient pursuits algorithm

作者：Xu Su^*; Yin Hongpeng; Tan Xue; Chai Yi

来源：Journal of Computational Information Systems, 2012, 8(7): 2973-2980.

• Compressive sensing
• Conjugate gradient
• Convex optimization problems
• Gradient pursuits
• Reconstructed image
• Sparse approximations

This paper presents a fast improved approximate conjugate gradient pursuits (IGP) algorithm for 2-D image compression. Firstly symmlet walvet function is employed to convert 2-D image to the sparse approximation matrix. Secondly an improved gradient pursuits algorithm (IGP) is presented to solve the convex optimization problem. Finally, reconstructed image is built up by symmlet walvet. Experimental results show that the proposed approach, compares with BP, MP, OMP and GP, has highest compression ratio and lowest computational complexity in the same image quality.

• 出版日期2012