Speaker linking using complete-linkage clustering

Speaker diarization determines instances of the same speaker within a recording. Extending this task to a collection of recordings for linking together segments spoken by a unique speaker requires speaker linking. In this paper we propose a speaker linking system using linkage clustering and state-of-the-art speaker recognition techniques. We evaluate our approach against two baseline linking systems using agglomerative cluster merging (AC) and agglomerative clustering with model retraining (ACR). We demonstrate that our linking method, using complete-linkage clustering, provides a relative improvement of 20% and 29% in attribution error rate (AER), over the AC and ACR systems, respectively.

Speaker attribution of multiple telephone conversations using a complete-linkage clustering approach

In this paper we propose and evaluate a speaker attribution system using a complete-linkage clustering method. Speaker attribution refers to the annotation of a collection of spoken audio based on speaker identities. This can be achieved using diarization and speaker linking. The main challenge associated with attribution is achieving computational efficiency when dealing with large audio archives. Traditional agglomerative clustering methods with model merging and retraining are not feasible for this purpose. This has motivated the use of linkage clustering methods without retraining. We first propose a diarization system using complete-linkage clustering and show that it outperforms traditional agglomerative and single-linkage clustering based diarization systems with a relative improvement of 40% and 68%, respectively. We then propose a complete-linkage speaker linking system to achieve attribution and demonstrate a 26% relative improvement in attribution error rate (AER) over the single-linkage speaker linking approach.

Complete-linkage clustering for voice activity detection in audio and visual speech

We propose a novel technique for conducting robust voice activity detection (VAD) in high-noise recordings. We use Gaussian mixture modeling (GMM) to train two generic models; speech and non-speech. We then score smaller segments of a given (unseen) recording against each of these GMMs to obtain two respective likelihood scores for each segment. These scores are used to compute a dissimilarity measure between pairs of segments and to carry out complete-linkage clustering of the segments into speech and non-speech clusters. We compare the accuracy of our method against state-of-the-art and standardised VAD techniques to demonstrate an absolute improvement of 15% in half-total error rate (HTER) over the best performing baseline system and across the QUT-NOISE-TIMIT database. We then apply our approach to the Audio-Visual Database of American English (AVDBAE) to demonstrate the performance of our algorithm in using visual, audio-visual or a proposed fusion of these features.

Extending the task of diarization to speaker attribution

In this paper we extend the concept of speaker annotation within a single-recording, or speaker diarization, to a collection wide approach we call speaker attribution. Accordingly, speaker attribution is the task of clustering expectantly homogenous intersession clusters obtained using diarization according to common cross-recording identities. The result of attribution is a collection of spoken audio across multiple recordings attributed to speaker identities. In this paper, an attribution system is proposed using mean-only MAP adaptation of a combined-gender UBM to model clusters from a perfect diarization system, as well as a JFA-based system with session variability compensation. The normalized cross-likelihood ratio is calculated for each pair of clusters to construct an attribution matrix and the complete linkage algorithm is employed to conduct clustering of the inter-session clusters. A matched cluster purity and coverage of 87.1% was obtained on the NIST 2008 SRE corpus.

Robust automatic speaker linking and attribution

This research makes a major contribution which enables efficient searching and indexing of large archives of spoken audio based on speaker identity. It introduces a novel technique dubbed as “speaker attribution” which is the task of automatically determining ‘who spoke when?’ in recordings and then automatically linking the unique speaker identities within each recording across multiple recordings. The outcome of the research will also have significant impact in improving the performance of automatic speech recognition systems through the extracted speaker identities.

Speaker attribution of Australian broadcast news data

Speaker attribution is the task of annotating a spoken audio archive based on speaker identities. This can be achieved using speaker diarization and speaker linking. In our previous work, we proposed an efficient attribution system, using complete-linkage clustering, for conducting attribution of large sets of two-speaker telephone data. In this paper, we build on our proposed approach to achieve a robust system, applicable to multiple recording domains. To do this, we first extend the diarization module of our system to accommodate multi-speaker (>2) recordings. We achieve this through using a robust cross-likelihood ratio (CLR) threshold stopping criterion for clustering, as opposed to the original stopping criterion of two speakers used for telephone data. We evaluate this baseline diarization module across a dataset of Australian broadcast news recordings, showing a significant lack of diarization accuracy without previous knowledge of the true number of speakers within a recording. We thus propose applying an additional pass of complete-linkage clustering to the diarization module, demonstrating an absolute improvement of 20% in diarization error rate (DER). We then evaluate our proposed multi-domain attribution system across the broadcast news data, demonstrating achievable attribution error rates (AER) as low as 17%.

A speaker rediarization scheme for improving diarization in large two-speaker telephone datasets

In this paper we propose a novel scheme for carrying out speaker diarization in an iterative manner. We aim to show that the information obtained through the first pass of speaker diarization can be reused to refine and improve the original diarization results. We call this technique speaker rediarization and demonstrate the practical application of our rediarization algorithm using a large archive of two-speaker telephone conversation recordings. We use the NIST 2008 SRE summed telephone corpora for evaluating our speaker rediarization system. This corpus contains recurring speaker identities across independent recording sessions that need to be linked across the entire corpus. We show that our speaker rediarization scheme can take advantage of inter-session speaker information, linked in the initial diarization pass, to achieve a 30% relative improvement over the original diarization error rate (DER) after only two iterations of rediarization.

Common variants in the CYP2C19 gene are associated with susceptibility to endometriosis

Objective: To follow-up previous studies highlighting a possible role for cytochrome P450, family 2, subfamily C, 19 (CYP2C19) in susceptibility to endometriosis by searching for additional variants in the CYP2C19 gene that may be associated with the disease. Design Case-control study. Setting Academic research. Subject(s) The cases comprised 2,271 women with surgically confirmed endometriosis; the controls comprised 939 women with self-report of no endometriosis and 1,770 unscreened population samples. Intervention(s) Sequencing of the CYP2C19 region and follow-up of 80 single nucleotide polymorphisms (SNPs) in two case-control samples. Main Outcome Measure(s) Allele frequency differences between cases and controls. Result(s) Sequencing of the CYP2C19 gene region resulted in the detection of a large number of known and novel SNPs. Genotyping of 80 polymorphic SNPs in 901 endometriosis cases and 939 controls resulted in study-wide significant association signals for SNPs in moderate or complete linkage disequilibrium with rs4244285, a functional SNP in exon 5 that abrogates CYP2C19 function through the creation of an alternative splice site. Evidence of association was also detected for another functional SNP in the CYP2C19 promoter, rs12248560, which was highlighted in our previous study. Conclusion(s) Functional variants in CYP2C19 may contribute to endometriosis susceptibility in both familial and sporadic cases. © 2014 by American Society for Reproductive Medicine.

A cluster-voting approach for speaker diarization and linking of Australian broadcast news recordings

We present a clustering-only approach to the problem of speaker diarization to eliminate the need for the commonly employed and computationally expensive Viterbi segmentation and realignment stage. We use multiple linear segmentations of a recording and carry out complete-linkage clustering within each segmentation scenario to obtain a set of clustering decisions for each case. We then collect all clustering decisions, across all cases, to compute a pairwise vote between the segments and conduct complete-linkage clustering to cluster them at a resolution equal to the minimum segment length used in the linear segmentations. We use our proposed cluster-voting approach to carry out speaker diarization and linking across the SAIVT-BNEWS corpus of Australian broadcast news data. We compare our technique to an equivalent baseline system with Viterbi realignment and show that our approach can outperform the baseline technique with respect to the diarization error rate (DER) and attribution error rate (AER).

Linkage disequilibrium analysis in the genetically isolated Norfolk Island population

Norfolk Island is a human genetic isolate, possessing unique population characteristics that could be utilized for complex disease gene localization. Our intention was to evaluate the extent and strength of linkage disequilibrium (LD) in the Norfolk isolate by investigating markers within Xq13.3 and the NOS2A gene encoding the inducible nitric oxide synthase. A total of six microsatellite markers spanning approximately 11 Mb were assessed on chromosome Xq13.3 in a group of 56 men from Norfolk Island. Additionally, three single nucleotide polymorphisms (SNPs) localizing to the NOS2A gene were analyzed in a subset of the complex Norfolk pedigree. With the exception of two of the marker pairs, one of which is the most distantly spaced marker, all the Xq13.3 marker pairs were found to be in significant LD indicating that LD extends up to 9.5-11.5 Mb in the Norfolk Island population. Also, all SNPs studied showed significant LD in both Norfolk Islanders and Australian Caucasians, with two of the marker pairs in complete LD in the Norfolk population only. The Norfolk Island study population possesses a unique set of characteristics including founder effect, geographical isolation, exhaustive genealogical information and phenotypic data of use to cardiovascular disease risk traits. With LD extending up to 9.5-11 Mb, the Norfolk isolate should be a powerful resource for the localization of complex disease genes.

Association and linkage analyses of restriction fragment length polymorphisms for the human renin and antithrombin III genes in essential hypertension

Essential hypertension is a highly hereditable disorder in which genetic influences predominate over environmental factors. The molecular genetic profiles which predispose to essential hypertension are not known. In rats with genetic hypertension, there is some recent evidence pointing to linkage of renin gene alleles with blood pressure. The genes for renin and antithrombin III belong to a conserved synteny group which, in humans, spans the q21.3-32.3 region of chromosome I and, in rats, is linkage group X on chromosome 13. The present study examined the association of particular human renin gene (REN) and antithrombin III gene (AT3) polymorphisms with essential hypertension by comparing the frequency of specific alleles for each of these genes in 50 hypertensive offspring of hypertensive parents and 91 normotensive offspring of normotensive parents. In addition, linkage relationships were examined in hypertensive pedigrees with multiple affected individuals. Alleles of a REN HindIII restriction fragment length polymorphism (RFLP) were detected using a genomic clone, λHR5, to probe Southern blots of HindIII-cut leucocyte DNA, and those for an AT3 Pstl RFLP were detected by phATIII 113 complementary DNA probe. The frequencies of each REN allele in the hypertensive group were 0.76 and 0.24 compared with 0.74 and 0.26 in the normotensive group. For AT3, hypertensive allele frequencies were 0.49 and 0.51 compared with normotensive values of 0.54 and 0.46. These differences were not significant by χ2 analysis (P > 0.2). Linkage analysis of a family (data from 16 family members, 10 of whom were hypertensive), informative for both markers, without an age-of-onset correction, and assuming dominant inheritance of hypertension, complete penetrance and a disease frequency of 20%, did not indicate linkage of REN with hypertension, but gave a positive, although not significant, logarithm of the odds for linkage score of 0.784 at a recombination fraction of 0 for AT3 linkage to hypertension. In conclusion, the present study could find no evidence for an association of a REN HindIII RFLP with essential hypertension or for a linkage of the locus defined by this RFLP in a family segregating for hypertension. In the case of an AT3 Pstl RFLP, although association analysis was negative, linkage analysis suggested possible involvement (odds of 6:1 in favour) of a gene located near the 1q23 locus with hypertension in one informative family.