Fig. 11From: Learning-based robust speaker counting and separation with the aid of spatial coherenceConfusion matrices for the speaker counting results obtained using a baseline 1, b baseline 2, c proposal 1, and d proposal 2Back to article page