Search Results for author:
Found 18 papers, 2 papers with code
Overlaps and Gender Analysis in the Context of Broadcast Media
Our main goal is to study the interactions between speakers according to their gender and role in broadcast media.
Channel-Combination Algorithms for Robust Distant Voice Activity and Overlapped Speech Detection
A channel-number invariant loss is proposed to learn a unique feature representation regardless of the number of available microphones.
Joint speech and overlap detection: a benchmark over multiple audio setup and speech domains
Voice activity and overlapped speech detection (respectively VAD and OSD) are key pre-processing tasks for speaker diarization.
Multi-microphone Automatic Speech Segmentation in Meetings Based on Circular Harmonics Features
Pipeline systems rely on speech segmentation to extract speakers' segments and achieve robust speaker diarization.
Evaluation of Speaker Anonymization on Emotional Speech
This paper studies the impact of the speaker anonymization baseline system of the VPC on emotional information present in speech utterances.
Are disentangled representations all you need to build speaker anonymization systems?
We propose enhancing the disentanglement by removing speaker information from the acoustic model using vector quantization.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+4
Privacy-Preserving Speech Representation Learning using Vector Quantization
With the popularity of virtual assistants (e. g., Siri, Alexa), the use of speech recognition is now becoming more and more widespread. However, speech signals contain a lot of sensitive information, such as the speaker's identity, which raises privacy concerns. The presented experiments show that the representations extracted by the deep layers of speech recognition networks contain speaker information. This paper aims to produce an anonymous representation while preserving speech recognition performance. To this end, we propose to use vector quantization to constrain the representation space and induce the network to suppress the speaker identity. The choice of the quantization dictionary size allows to configure the trade-off between utility (speech recognition) and privacy (speaker identity concealment).
On the invertibility of a voice privacy system using embedding alignement
This paper explores various attack scenarios on a voice anonymization system using embeddings alignment techniques.
Evaluating X-vector-based Speaker Anonymization under White-box Assessment
In the scenario of the Voice Privacy challenge, anonymization is achieved by converting all utterances from a source speaker to match the same target identity; this identity being randomly selected.
A Study of F0 Modification for X-Vector Based Speech Pseudonymization Across Gender
Speech pseudonymization aims at altering a speech signal to map the identifiable personal characteristics of a given speaker to another identity.
End-to-end anti-spoofing with RawNet2
Spoofing countermeasures aim to protect automatic speaker verification systems from attempts to manipulate their reliability with the use of spoofed speech signals.
\'Evaluation de syst\`emes apprenant tout au long de la vie (Evaluation of lifelong learning systems )
Une adaptation de leur mod{\`e}le par des experts en apprentissage automatique est possible mais tr{\`e}s co{\^u}teuse alors que les soci{\'e}t{\'e}s utilisant ces syst{\`e}mes disposent d{'}experts du domaine qui pourraient accompagner ces syst{\`e}mes dans un apprentissage tout au long de la vie.
Evaluation of Lifelong Learning Systems
Current intelligent systems need the expensive support of machine learning experts to sustain their performance level when used on a daily basis.
I4U Submission to NIST SRE 2018: Leveraging from a Decade of Shared Experiences
The I4U consortium was established to facilitate a joint entry to NIST speaker recognition evaluations (SRE).
Exploration de param\`etres acoustiques d\'eriv\'es de GMM pour l'adaptation non supervis\'ee de mod\`eles acoustiques \`a base de r\'eseaux de neurones profonds (Exploring GMM-derived features for unsupervised adaptation of deep neural network acoustic models)
L{'}{\'e}tude pr{\'e}sent{\'e}e dans cet article am{\'e}liore une m{\'e}thode r{\'e}cemment propos{\'e}e pour l{'}adaptation de mod{\`e}les acoustiques markoviens coupl{\'e}s {\`a} un r{\'e}seau de neurones profond (DNN-HMM).
Autoapprentissage pour le regroupement en locuteurs : premi\`eres investigations (First investigations on self trained speaker diarization )
This paper investigates self trained cross-show speaker diarization applied to collections of French TV archives, based on an \textit{i-vector/PLDA} framework.
Fantastic 4 system for NIST 2015 Language Recognition Evaluation
This article describes the systems jointly submitted by Institute for Infocomm (I$^2$R), the Laboratoire d'Informatique de l'Universit\'e du Maine (LIUM), Nanyang Technology University (NTU) and the University of Eastern Finland (UEF) for 2015 NIST Language Recognition Evaluation (LRE).
