A very fast large scale BSS algorithm by joint approximate diagonalization of simplified cumulant matrices
首发时间:2007-12-17
Abstract:This paper works on joint approximate diagonalization of simplified fourth order cumulant matrices for very fast and large scale blind separation of instantaneous mixing model sources. The JADE algorithm is widely accepted but only limited to small scale separation tasks. The SHIBBS algorithm calculates a fraction of the fourth order cumulant set and avoids eigenmatrix decomposition to reduce calculation cost. However it was seen to be slower than JADE at the time of its first publication and is hence less known. On the other hand, the SJAD algorithm using the same approach is shown to be very fast. In this paper the cumulant matrices are further simplified, a new iteration convergence criterion is proposed and a unified optimized iteration threshold is determined. Both theoretical analysis and experimental separation proves that this algorithm is memory efficient, as robust and reliable as and very much faster than JADE. Hence it is suitable for both small and large scale blind source separation problems.
点击查看论文中文信息
A very fast large scale BSS algorithm by joint approximate diagonalization of simplified cumulant matrices
摘要:This paper works on joint approximate diagonalization of simplified fourth order cumulant matrices for very fast and large scale blind separation of instantaneous mixing model sources. The JADE algorithm is widely accepted but only limited to small scale separation tasks. The SHIBBS algorithm calculates a fraction of the fourth order cumulant set and avoids eigenmatrix decomposition to reduce calculation cost. However it was seen to be slower than JADE at the time of its first publication and is hence less known. On the other hand, the SJAD algorithm using the same approach is shown to be very fast. In this paper the cumulant matrices are further simplified, a new iteration convergence criterion is proposed and a unified optimized iteration threshold is determined. Both theoretical analysis and experimental separation proves that this algorithm is memory efficient, as robust and reliable as and very much faster than JADE. Hence it is suitable for both small and large scale blind source separation problems.
论文图表:
引用
No.1701317725511978****
同行评议
共计0人参与
勘误表
A very fast large scale BSS algorithm by joint approximate diagonalization of simplified cumulant matrices
评论
全部评论0/1000