2014 Volume E97.A Issue 4 Pages 1024-1026
Canonical correlation analysis (CCA) is applied to extract features for microphone classification. We utilized the coherence between near-silence regions. Experimental results show the promise of canonical correlation features for microphone classification.