B. Adaptive multiple wave subtraction using independent component analysis
基于独立分量分析的多次波自适应相减技术

We present a new adaptive multiple subtraction algorithm based on independent component analysis (ICAAMS). The output energy minimum criterion, which is adopted by almost all the existing adaptive multiple subtraction methods, is based on second order statistics. The method proposed in this paper maximizes the nongaussianity of the output signal, which can be measured by higher order statistics. The method has been applied to several synthetic datasets generated by simple convolution and finite-difference model (FDM) technique, and the results show that our method can separate primary wave and multiple wave effectively.