Text-Independent Speaker Verification Based on Deep Neural Networks and Segmental Dynamic Time Warping | IEEE Conference Publication | IEEE Xplore