压缩感知矩阵:二值与三值对比
信息论
2014-02-10 v5 math.IT
摘要
二值矩阵和三值矩阵是压缩感知中两种流行的感知矩阵,因其具有竞争力的性能和低计算量而备受青睐。然而,据我们所知,尽管这在实践中很重要,但似乎尚无文献旨在评估它们在具有相同稀疏度时的性能。基于 RIP 分析和数值模拟,本文首次揭示,如果{0, 1}二值矩阵与{0, +1, -1}三值矩阵在非零位置上的分布相同,则前者具有更好的整体性能。
引用
@article{arxiv.1304.4161,
title = {Compressed Sensing Matrices: Binary vs. Ternary},
author = {Weizhi Lu and Kidiyo Kpalma and Joseph Ronsin},
journal= {arXiv preprint arXiv:1304.4161},
year = {2014}
}
备注
This paper has been withdrawn by the authors. The proof is irrigorous so that the conclusion is not completely right