Figure 13From: A sub-band-based feature reconstruction approach for robust speaker recognition The recognition performance derived from different division ways. Back to article page