Fig. 2From: Text-independent speaker recognition based on adaptive course learning loss and deep residual networkFlow chart of residual block constructionBack to article page