Speaker diarization with i-vectors from DNN senone posteriors
نویسندگان
چکیده
Motivated by recent gains in speaker identification by incorporating senone posteriors from deep neural networks (DNNs) into i-vector extraction, we examine similar enhancements to speaker diarization with i-vector clustering. We examine two DNNs with different numbers of senone targets in combination with a diagonal or full covariance universal background model (UBM) in the context of the multilingual corpus CALLHOME. Results show that the larger DNN with a full covariance UBM gives the best performance. The improvements appear to have a strong dependence on number of speakers in a conversation, and a lesser dependence on language. Overall, when combined with resegmentation, the proposed system improves CALLHOME performance to 10.3% DER.
منابع مشابه
Text-Available Speaker Recognition System for Forensic Applications
This paper examines a text-available speaker recognition approach targeting scenarios where the transcripts of test utterances are either available or obtainable through manual transcription. Forensic speaker recognition is one of such applications where the human supervision can be expected. In our study, we extend an existing Deep Neural Network (DNN) ivector-based speaker recognition system ...
متن کاملDNN based Speaker Recognition on Short Utterances
This paper investigates the effects of limited speech data in the context of speaker verification using deep neural network (DNN) approach. Being able to reduce the length of required speech data is important to the development of speaker verification system in real world applications. The experimental studies have found that DNN-senone-based Gaussian probabilistic linear discriminant analysis ...
متن کاملExploiting Eigenposteriors for Semi-Supervised Training of DNN Acoustic Models with Sequence Discrimination
Deep neural network (DNN) acoustic models yield posterior probabilities of senone classes. Recent studies support the existence of low-dimensional subspaces underlying senone posteriors. Principal component analysis (PCA) is applied to identify eigenposteriors and perform low-dimensional projection of the training data posteriors. The resulted enhanced posteriors are applied as soft targets for...
متن کاملInvestigating factor analysis features for deep neural networks in noisy speech recognition
The problem of speaker and channel adaptation in deep neural network (DNN) based automatic speech recognition (ASR) systems is of substantial interest in advancing the performance of these systems. Recently, the speaker identity vectors (i-vectors) have shown improvements for ASR systems in matched conditions. In this paper, we propose the application of the general factor analysis framework fo...
متن کاملDNN i-Vector Speaker Verification with Short, Text-Constrained Test Utterances
We investigate how to improve the performance of DNN ivector based speaker verification for short, text-constrained test utterances, e.g. connected digit strings. A text-constrained verification, due to its smaller, limited vocabulary, can deliver better performance than a text-independent one for a short utterance. We study the problem with “phonetically aware” Deep Neural Net (DNN) in its cap...
متن کامل