A Separation method of one-channel heart voice signal

This paper presents a new BSS algorithm of single mixed heart voice signal. First the paper discussed the separation model of single-channel heart voice signal and the layered principles. How does this layered work? Can be separated into many layers? The problems are found the solution by to add a reference signal. The equal length layered technique and how to gain independent sub-band functions are studied. Enter the single-channel mixed heart voice signal through the independent sub-band functions, causes it be transformed into the multi-dimensional vector by a dimensional vector, and realized its blind separation and the algorithm does not use the prior knowledge. Finally through the heart voice separation experiments confirmed this method validity and the feasibility.