Speaker Diarization for Multi-microphone Meetings Using Only Between-Channel Differences

Pardo, José Manuel; Anguera, Xavier; Wooters, Chuck

doi:10.1007/11965152_23

Cited by 30 publications

(33 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Nonspeech frames, estimated previously, are excluded from the subsequent process. The vector of delays is then fed into the aforementioned segmentation and agglomerative clustering module instead of the acoustic vectors [36]. We experimented with several values for the segmentation and agglomerative clustering parameters such as the initial number of mixtures per cluster and the number of initial clusters as mentioned in the previous section.…”

Section: Baselinementioning

confidence: 99%

Speaker Diarization For Multiple-Distant-Microphone Meetings Using Several Sources of Information

Pardo

Anguera

Wooters

2007

IEEE Trans. Comput.

Self Cite

View full text Add to dashboard Cite

Abstract-Human-machine interaction in meetings requires the localization and identification of the speakers interacting with the system, as well as the recognition of the words spoken. A seminal step toward this goal is the field of rich transcription research, which includes speaker diarization together with the annotation of sentence boundaries and the elimination of speaker disfluencies. The subarea of speaker diarization attempts to identify the number of participants in a meeting and create a list of speech time intervals for each such participant. In this paper, we analyze the correlation between signals coming from multiple microphones and propose an improved method for carrying out speaker diarization for meetings with multiple distant microphones. The proposed algorithm makes use of acoustic information and information from the delays between signals coming from the different sources. Using this procedure, we were able to achieve state-of-the-art performance in the NIST spring 2006 rich transcription evaluation, improving the Diarization Error Rate (DER) by 15 percent to 28 percent relative to previous systems.

show abstract

Section: Baselinementioning

confidence: 99%

Speaker Diarization For Multiple-Distant-Microphone Meetings Using Several Sources of Information

Pardo

Anguera

Wooters

2007

IEEE Trans. Comput.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Experiments have shown that as stand alone features [7], TDOA performs poorly respect to MFCC but significant performance improvements are obtained when TDOA are used in combination with MFCC [8], [9].…”

mentioning

confidence: 99%

An Information Theoretic Combination of MFCC and TDOA Features for Speaker Diarization

Vijayasenan

Valente

Bourlard

2011

IEEE Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

This correspondence describes a novel system for speaker diarization of meetings recordings based on the combination of acoustic features (MFCC) and Time Delay of Arrivals (TDOA). The first part of the paper analyzes differences between MFCC and TDOA features which possess completely different statistical properties. When Gaussian Mixture Models are used, experiments reveal that the diarization system is sensitive to the different recording scenarios (i.e. meeting rooms with varying number of microphones). In the second part, a new multistream diarization system is proposed extending previous work on Information Theoretic diarization. Both speaker clustering and speaker realignment steps are discussed; in contrary to current systems, the proposed method avoids to perform the feature combination averaging log-likelihood scores. Experiments on meetings data reveal that the proposed approach outperforms the GMM based system when the recording is done with varying number of microphones.

show abstract

“…Conventional speaker diarization systems [1] use an ergodic Hidden Markov Model (HMM) with speakers as HMM states. Good results were achieved by the systems using the combination of MelFrequency Cepstral Coefficients (MFCC) and Time Difference of Arrival (TDOA) features [2] with arrays composed of different number of microphones, while performance of standalone TDOA features was estimated as poor in respect to MFCC [3]. TDOA features can be used without prior knowledge of the geometry of the microphone array.…”

Section: Introductionmentioning

confidence: 99%

Audio spatio-temporal fingerprints for cloudless real-time hands-free diarization on mobile devices

Korchagin

2011

2011 Joint Workshop on Hands-Free Speech Communication and Microphone Arrays

View full text Add to dashboard Cite

In this paper, we propose a new low bit rate representation of a sound field and a new method for the corresponding cloudless low delay hands-free diarization suitable for lowperformance mobile devices, e.g. mobile phones. The proposed audio spatio-temporal fingerprint representation results in low bit rate (500 bytes/second), however contains complete information about continuous audio tracking of multiple acoustic sources in an open, unconstrained environment. The core of the algorithm is based on simultaneous multiple data stream processing using audio spatio-temporal fingerprint representation to cover higher level events relevant for diarization, e.g. turns, interruptions, crosstalk, speech and non-speech segments. Performance levels achieved to date on 5 hours of hand-labelled datasets have shown the feasibility of the approach at the same time as resulting in 7.58% CPU load on 1-core ultra-low-power mobile processor running at 1 GHz and low algorithmic delay of 112 ms.

show abstract

Speaker Diarization for Multi-microphone Meetings Using Only Between-Channel Differences

Cited by 30 publications

References 6 publications

Speaker Diarization For Multiple-Distant-Microphone Meetings Using Several Sources of Information

Speaker Diarization For Multiple-Distant-Microphone Meetings Using Several Sources of Information

An Information Theoretic Combination of MFCC and TDOA Features for Speaker Diarization

Audio spatio-temporal fingerprints for cloudless real-time hands-free diarization on mobile devices

Contact Info

Product

Resources

About