Feature Smoothing and Frame Reduction for Speaker Recognition | IEEE Conference Publication | IEEE Xplore