Deep Normalization for Speaker Vectors
This work addresses a specific bottleneck in speaker recognition for applications like biometrics, but it is incremental as it builds on existing deep embedding methods.
The paper tackles the problem of non-Gaussian and non-homogeneous distributions in deep speaker vectors, which degrade speaker recognition performance, and proposes a deep normalization approach using a discriminative normalization flow model, achieving substantial performance gains in experiments on SITW and CNCeleb corpora.
Deep speaker embedding has demonstrated state-of-the-art performance in speaker recognition tasks. However, one potential issue with this approach is that the speaker vectors derived from deep embedding models tend to be non-Gaussian for each individual speaker, and non-homogeneous for distributions of different speakers. These irregular distributions can seriously impact speaker recognition performance, especially with the popular PLDA scoring method, which assumes homogeneous Gaussian distribution. In this paper, we argue that deep speaker vectors require deep normalization, and propose a deep normalization approach based on a novel discriminative normalization flow (DNF) model. We demonstrate the effectiveness of the proposed approach with experiments using the widely used SITW and CNCeleb corpora. In these experiments, the DNF-based normalization delivered substantial performance gains and also showed strong generalization capability in out-of-domain tests.