No abstract

Infeasible interior point methods have been very popular and effective. In this paper, we propose a predictor-corrector infeasible interior point algorithm for convex quadratic programming, and we prove its convergence and analyze its complexity. The algorithm has the polynomial numerical complexity with O(nL)-iteration.

No abstract

According to the different characteristics that signal and noise exhibit during wavelet decomposition, a new denoising method based on the lifting scheme wavelet packet decomposition is presented. In this method, the SAR images are decomposed by using the best wavelet packet and the norm of each sub-band are calculated; signals and noise can be discriminated based on the norm and soft-threshold method, and the images can be denoised. Experiments show that the proposed algorithm has excellent performance in denoising SAR images, and can remove most noise of images with well-kept texture detail information. The calculating speed of the method is twice the speed of the general wavelet packet transform algorithm.

No abstract