Text-Independent Speaker Verification Using 3D Convolutional Neural Networks

In this paper, a novel method using 3D Convolutional Neural Network (3D-CNN) architecture has been proposed for speaker verification in the text-independent setting. At the development phase, a CNN is trained to classify speakers at the utterance-level. In the enrollment stage, the trained network is utilized to directly create a speaker model for each speaker based on the extracted features. Finally, in the evaluation phase, the extracted features from the test utterance will be compared to the stored speaker model to verify the claimed identity. One of the main challenges is the creation of the speaker models. Previouslyreported approaches create speaker models based on averaging the extracted features from utterances of the speaker, which is known as a d-vector system. In our paper, we propose an adaptive feature learning by utilizing the 3D-CNNs for direct speaker model creation in which, for both development and enrollment phases, an identical number of spoken utterances per speaker is fed to the network for representing the speaker utterances and creation of the speaker model. This leads to simultaneously capturing the speaker-related information and building a more robust system to cope with within-speaker variation. We demonstrate that the proposed method significantly outperforms the traditional d-vector verification system. Moreover, we empirically show that the proposed method is more effective than the endto-end learning method.