Spoofing Attack Detection using the Non-linear Fusion of Sub-band Classifiers

Authors: Hemlata Tak, Jose Patino, Andreas Nautsch, Nicholas Evans, Massimiliano Todisco

Published: 2020-05-20 23:37:28+00:00

Comment: Submitted to Interspeech 2020 conference, 5 pages

AI Summary

This paper proposes a spoofing attack detection system that leverages a bank of simple sub-band classifiers, each tuned to different spoofing attacks, combined via non-linear score fusion. This approach demonstrates superior performance, outperforming most sophisticated ensemble solutions relying on complex neural networks. The method achieved competitive results in the ASVspoof 2019 challenge, surpassing all but two of the 48 submitted systems for the logical access condition.

Abstract

The threat of spoofing can pose a risk to the reliability of automatic speaker verification. Results from the bi-annual ASVspoof evaluations show that effective countermeasures demand front-ends designed specifically for the detection of spoofing artefacts. Given the diversity in spoofing attacks, ensemble methods are particularly effective. The work in this paper shows that a bank of very simple classifiers, each with a front-end tuned to the detection of different spoofing attacks and combined at the score level through non-linear fusion, can deliver superior performance than more sophisticated ensemble solutions that rely upon complex neural network architectures. Our comparatively simple approach outperforms all but 2 of the 48 systems submitted to the logical access condition of the most recent ASVspoof 2019 challenge.

Key findings

The study found that non-linear fusion methods (GMM-based and SVM) significantly outperform linear fusion for combining scores from sub-band classifiers. Despite its simplicity, the proposed system, based on sub-band LFCC features and GMM back-ends with non-linear fusion, achieved performance comparable to state-of-the-art complex neural network solutions, outperforming all but two systems in the ASVspoof 2019 challenge. High-resolution front-ends and sub-band specific analysis are crucial for effective spoofing detection.

Approach

The approach involves an ensemble of simple classifiers, each equipped with a Linear Frequency Cepstral Coefficient (LFCC) front-end adapted to specific sub-bands identified using a Centre-of-Mass function, and a Gaussian Mixture Model (GMM) back-end. Scores from these individual sub-band classifiers (along with a full-band classifier) are then combined using non-linear fusion techniques, such as GMM-based fusion or Support Vector Machines (SVM).

Datasets

ASVspoof 2019 LA database

Model(s)

LFCC (Linear Frequency Cepstral Coefficients) front-ends, GMM (Gaussian Mixture Model) back-end, GMM-based fusion, SVM (Support Vector Machine) with polynomial kernel, Multinomial Logistic Regression for fusion.

Author countries

France

← Previous