Long Term Spectral Statistics for Voice Presentation Attack Detection

Automatic speaker verification systems can be spoofed through recorded, synthetic or voice converted speech of target speakers. To make these systems practically viable, the detection of such attacks, referred to as presentation attacks, is of paramount interest. In that direction, this paper investigates two aspects: (a) a novel approach to detect presentation attacks where, unlike conventional approaches, no speech signal related assumptions are made, rather the attacks are detected by computing first order and second order spectral statistics and feeding them to a classifier, and (b) generalization of the presentation attack detection systems across databases. Our investigations on Interspeech 2015 ASVspoof challenge dataset and AVspoof dataset show that, when compared to the approaches based on conventional short-term spectral processing, the proposed approach with a linear discriminative classifier yields a better system, irrespective of whether the spoofed signal is replayed to the microphone or is directly injected into the system software process. Cross-database investigations show that neither the short-term spectral processing based approaches nor the proposed approach yield systems which are able to generalize across databases or methods of attack. Thus, revealing the difficulty of the problem and the need for further resources and research.

Long-Term Spectral Statistics for Voice Presentation Attack Detection

Sébastien Marcel, Hannah Muckenhirn

Automatic speaker verification systems can be spoofed through recorded, synthetic or voice converted speech of target speakers. To make these systems practically viable, the detection of such attacks, referred to as presentation attacks, is of paramount interest. In that direction, this paper investigates two aspects: (a) a novel approach to detect presentation attacks where, unlike conventional approaches, no speech signal modeling related assumptions are made, rather the attacks are detected by computing first order and second order spectral statistics and feeding them to a classifier, and (b) generalization of the presentation attack detection systems across databases. Our investigations on ASVspoof 2015 challenge database and AVspoof database show that, when compared to the approaches based on conventional short-term spectral features, the proposed approach with a linear discriminative classifier yields a better system, irrespective of whether the spoofed signal is replayed to the microphone or is directly injected into the system software process. Cross-database investigations show that neither the short-term spectral processing based approaches nor the proposed approach yield systems which are able to generalize across databases or methods of attack. Thus, revealing the difficulty of the problem and the need for further resources and research.

2017

Presentation attack detection in voice biometrics

Sébastien Marcel

Recent years have shown an increase in both the accuracy of biometric systems and their practical use. The application of biometrics is becoming widespread with fingerprint sensors in smartphones, automatic face recognition in social networks and video-based applications, and speaker recognition in phone banking and other phone-based services. The popularization of the biometric systems, however, exposed their major flaw --- high vulnerability to spoofing attacks. A fingerprint sensor can be easily tricked with a simple glue-made mold, a face recognition system can be accessed using a printed photo, and a speaker recognition system can be spoofed with a replay of pre-recorded voice. The ease with which a biometric system can be spoofed demonstrates the importance of developing efficient anti-spoofing systems that can detect both known (conceivable now) and unknown (possible in the future) spoofing attacks. Therefore, it is important to develop mechanisms that can detect such attacks, and it is equally important for these mechanisms to be seamlessly integrated into existing biometric systems for practical and attack-resistant solutions. To be practical, however, an attack detection should have (i) high accuracy, (ii) be well-generalized for different attacks, and (iii) be simple and efficient. One reason for the increasing demand for effective presentation attack detection (PAD) systems is the ease of access to people's biometric data. So often, a potential attacker can almost effortlessly obtain necessary biometric samples from social networks, including facial images, audio and video recordings, and even extract fingerprints from high resolution images. Therefore, various privacy protection solutions, such as legal privacy requirements and algorithms for obfuscating personal information, e.g., visual privacy filters, as well as, social awareness of threats to privacy can also increase security of personal information and potentially reduce the vulnerability of biometric systems. In this chapter, however, we focus on presentation attacks detection in voice biometrics, i.e., automatic speaker verification (ASV) systems. We discuss vulnerabilities of these systems to presentation attacks (PAs), present different state of the art PAD systems, give the insights into their performances, and discuss the integration of PAD and ASV systems.

The Institution of Engineering and Technology2017

Boosting Localized Features for Speaker and Speech Recognition

Anindya Roy

In this thesis, we propose a novel approach for speaker and speech recognition involving localized, binary, data-driven features. The proposed approach is largely inspired by similar localized approaches in the computer vision domain. The success of these existing approaches coupled with their proven advantages of robustness and computational efficiency motivated us to apply these ideas to the speech domain. Our approach is distinct from the standard cepstral features-based approach for speaker and speech recognition. The proposed approach starts with a large set of simple localized features, each of which looks at very small parts of spectro-temporal representations of speech. Each feature is binary-valued. The most discriminative of these features are selected by boosting and combined to form the final classifier. Two systems are developed based on this general framework, a speaker recognition system and a speech recognition system. The speaker recognition system is evaluated under a wide range of experimental conditions, using clean speech, noisy speech and speech data collected from mobile phones. The system performs reliably in each condition, comparable with the standard systems using cepstral features and Gaussian Mixture Models. At the same time, it involves significantly lower number of floating point operations compared to these systems. In the case of the speech recognition system, we integrate our localized features with a Hidden Markov Model framework using multilayer perceptrons. Continuous speech recognition studies on standard databases show that these features perform equally well as cepstral features. It is also found that the fusion of these features with cepstral features leads to improved performance at both the feature level and the decision level. Apart from this, minor contributions include an audio-visual person recognition system developed using the same general approach of localized features described above, extending its applicability. Finally, a new (but related) class of localized features was developed for robust face detection.