Overlapping Speech Detection Using Long-Term Conversational Features for Speaker Diarization in Meeting Room Conversations

Yella, Sree Harsha; Bourlard, Hervé

doi:10.1109/taslp.2014.2346315

Cited by 37 publications

(19 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The current data segmentation approach makes measuring durations (or frequencies) of overlapping speech difficult. Previous research, however, showed that the number of turn switches is correlated with the number of interrupts and overlaps in a conversation [29,30]. We use this feature as a proxy for the amount of overlapping speech that occurs in a clinical interview.…”

Section: Dialogue Featuresmentioning

confidence: 99%

Identifying Mood Episodes Using Dialogue Features from Clinical Interviews

Aldeneh¹,

Jaiswal²,

Picheny³

et al. 2019

Interspeech 2019

View full text Add to dashboard Cite

Bipolar disorder, a severe chronic mental illness characterized by pathological mood swings from depression to mania, requires ongoing symptom severity tracking to both guide and measure treatments that are critical for maintaining long-term health. Mental health professionals assess symptom severity through semi-structured clinical interviews. During these interviews, they observe their patients' spoken behaviors, including both what the patients say and how they say it. In this work, we move beyond acoustic and lexical information, investigating how higher-level interactive patterns also change during mood episodes. We then perform a secondary analysis, asking if these interactive patterns, measured through dialogue features, can be used in conjunction with acoustic features to automatically recognize mood episodes. Our results show that it is beneficial to consider dialogue features when analyzing and building automated systems for predicting and monitoring mood.In contrast to previous work, the novelty of our work is three-fold: (1) we introduce a set of dialogue features to aid in the prediction of mood symptom severity; (2) we analyze dialogue features using a linear mixed effect model to study how mood episodes affect interaction patterns; (3) we show that explicitly adding high-level dialogue features to acoustic-based systems can improve the performance of automatic mood symptom severity prediction.

show abstract

Section: Dialogue Featuresmentioning

confidence: 99%

Identifying Mood Episodes Using Dialogue Features from Clinical Interviews

Aldeneh¹,

Jaiswal²,

Picheny³

et al. 2019

Interspeech 2019

View full text Add to dashboard Cite

show abstract

“…Overlapping speech occurs when there is more than one speaker speaking at any given instant of time in an audio recording. This is a very common phenomenon in spontaneous conversations like meeting room discussions, telephone conversations, television chat shows, and other similar media . Overlapping speech hinders the performance of speech processing systems such as the HMM based system, in two ways.…”

Section: Introductionmentioning

confidence: 99%

“…In ref. , the authors use features capturing higher level information from structure of a conversation such as silence and speaker change statistics to improve acoustic feature based classifier of overlapping and single‐speaker speech classes.…”

Section: Introductionmentioning

confidence: 99%

A hybrid method for overlapping speech detection in classroom environment

Jiang

Chen

Garg

2017

Comp Applic In Engineering

View full text Add to dashboard Cite

Classroom discourse, which is a major composition of classroom, contains a slew of useful information such as providing the feedback to the teachers may lead to an improvement of the teaching quality. Generally, the classroom discourse can be divided into four categories/activities: teacher discourse, student discourse, quiet, and discussion. The automatic classification of activities provides a practical way to deal with the classroom discourse. However, the recognition of the discussion activity (a kind of overlapping speech) poses a great challenge to the treatment of the classroom discourse analysis. Therefore, in this paper, a new hybrid method based on the silence distribution and Independent Component Analysis (ICA) was proposed for the detection of overlapping speech. The results were found to be satisfactory and in good agreement with an experimental data. Based on these experimental results, the performance of the speaker segment of classroom event can be efficiently analyzed.

show abstract

“…In addition, the speaker's voice identification and verification [4,5,6] are becoming attractive features for user-specific services. To provide such services, speaker clustering [7,8] plays a key role in identifying the number of speakers and grouping the utterances from the same user for the automatic user-specific model generation or speaker diarization [9,10].…”

Section: Introductionmentioning

confidence: 99%

Speaker Clustering by Iteratively Finding Discriminative Feature Space and Cluster Labels

2017

View full text Add to dashboard Cite

This paper presents a speaker clustering framework by iteratively performing two stages: a discriminative feature space is obtained given a cluster label set, and the cluster label set is updated using a clustering algorithm given the feature space. In the iterations of two stages, the cluster labels may be different from the true labels, and thus the obtained feature space based on the labels may be inaccurately discriminated. However, by iteratively performing above two stages, more accurate cluster labels and more discriminative feature space can be obtained, and finally they are converged. In this research, the linear discriminant analysis is used for discriminating the ivector feature space, and the variational Bayesian expectationmaximization on Gaussian mixture model is used for clustering the i-vectors. Our iterative clustering framework was evaluated using the database of keyword utterances and compared with the recently-published approaches. In all experiments, the results show that our framework outperforms the other approaches and converges in a few iterations.

show abstract

Overlapping Speech Detection Using Long-Term Conversational Features for Speaker Diarization in Meeting Room Conversations

Cited by 37 publications

References 35 publications

Identifying Mood Episodes Using Dialogue Features from Clinical Interviews

Identifying Mood Episodes Using Dialogue Features from Clinical Interviews

A hybrid method for overlapping speech detection in classroom environment

Speaker Clustering by Iteratively Finding Discriminative Feature Space and Cluster Labels

Contact Info

Product

Resources

About