Deep CNN based feature extractor for text-prompted speaker recognition

TL;DR

This paper explores the use of deep convolutional neural networks with Max-Feature-Map activation for text-prompted speaker verification, achieving state-of-the-art results with a novel multitask training approach.

Contribution

It introduces a deep CNN feature extractor with Max-Feature-Map activation and multitask learning for speaker verification, surpassing traditional methods.

Findings

Achieved 2.85% EER on RSR2015 dataset.

Max-Feature-Map acts as an embedded feature selector.

Fusion with baseline improves performance.

Abstract

Deep learning is still not a very common tool in speaker verification field. We study deep convolutional neural network performance in the text-prompted speaker verification task. The prompted passphrase is segmented into word states - i.e. digits -to test each digit utterance separately. We train a single high-level feature extractor for all states and use cosine similarity metric for scoring. The key feature of our network is the Max-Feature-Map activation function, which acts as an embedded feature selector. By using multitask learning scheme to train the high-level feature extractor we were able to surpass the classic baseline systems in terms of quality and achieved impressive results for such a novice approach, getting 2.85% EER on the RSR2015 evaluation set. Fusion of the proposed and the baseline systems improves this result.