摘要
低秩矩阵恢复算法主要包括鲁棒主成分分析、矩阵补全、低秩表示,由于矩阵补全是一个NP难的问题,低秩表示涉及到字典矩阵,复杂度高,因此本文主要针对鲁棒主成分分析在FPGA上的研究与应用进行了描述,并且在CPU以及FPGA上实现了图像恢复。实验结果表明,基于FPGA的HLS设计相对于传统CPU在速度上得到了数十倍的提高。
The low rank matrix recovery algorithm mainly includes the robust principal component analysis, the matrix completion, the low rank representation. Because the matrix completion is a NP hard problem, the low rank said relates to the dictionary matrix, high complexity. Therefore,this paper mainly described the research and application of the robust principal component analysis in FPGA,and the CPU and FPGA image restoration is realized. The experimental results show that the speed of the design based on the FPGA HLS has been increased by tens of times compared to the traditional CPU.
出处
《单片机与嵌入式系统应用》
2017年第2期11-14,共4页
Microcontrollers & Embedded Systems
