Fig. 12From: Learning-based robust speaker counting and separation with the aid of spatial coherenceGround truth speaker activities for a case I and b case IIBack to article page