Efficient audio segmentation algorithms based on the BIC

Cettolo, Mauro; Vescovi, Michele

doi:10.1109/icassp.2003.1201737

Cited by 37 publications

(37 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Alternatively, the generalized likelihood ratio (GLR) test can be applied [15,45]. The most popular criterion is BIC [7,8,11,36,37,40,41,44,[73][74][75][76]. Metric-based segmentation algorithms generally yield a high recall rate at a moderate precision rate.…”

Section: Speaker Segmentation Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Speaker segmentation and clustering

2008

View full text Add to dashboard Cite

This survey focuses on two challenging speech processing topics, namely: speaker segmentation and speaker clustering. Speaker segmentation aims at finding speaker change points in an audio stream, whereas speaker clustering aims at grouping speech segments based on speaker characteristics. Model-based, metric-based, and hybrid speaker segmentation algorithms are reviewed. Concerning speaker clustering, deterministic and probabilistic algorithms are examined. A comparative assessment of the reviewed algorithms is undertaken, the algorithm advantages and disadvantages are indicated, insight to the algorithms is offered, and deductions as well as recommendations are given. Rich transcription and movie analysis are candidate applications that benefit from combined speaker segmentation and clustering.

show abstract

Section: Speaker Segmentation Algorithmsmentioning

confidence: 99%

“…Moreover, BIC computations at the beginning of large analysis windows are ignored, since they would be repeated several times. The aforementioned heuristics, due to their efficiency, are commonly used by other researchers [41,44,51,73]. 24-order MFCCs are employed.…”

Section: Performance Of Bic-based Segmentation Algorithmsmentioning

confidence: 99%

Speaker segmentation and clustering

2008

View full text Add to dashboard Cite

show abstract

“…Recent research on audio segmentation mostly focused on four categories: energy based, model-based (Kemp et al, 2000), metric-based , and information criterion-based approaches (Cettolo and Federico, 2000;Cettolo and Vescovi, 2003;Chen and Gopalakrishnan, 1998). Energy audio segmentation only detects change-points at silence segments, which generally are not directly connected with the acoustic changes of the audio signals.…”

Section: Previous Workmentioning

confidence: 99%

Robust Unsupervised Speaker Segmentation for Audio Diarization

Kadri¹,

Davy²,

Ellouze³

2010

Signal Processing

View full text Add to dashboard Cite

“…These distances are then used to drive an agglomerative hierarchical speaker clustering based on the BIC stopping criterion to reduce the number of clusters by merging. The expression of ΔBIC is given by [7]:…”

Section: Speaker Clustering Via Proposed Apahcmentioning

confidence: 99%

Robust Speaker Clustering Using Affinity Propagation

Zhang

Lü

Suo

et al. 2008

IEICE Transactions on Information and Systems

View full text Add to dashboard Cite

SUMMARYIn this letter, a recently proposed clustering algorithm named affinity propagation is introduced for the task of speaker clustering. This novel algorithm exhibits fast execution speed and finds clusters with low error. However, experiments show that the speaker purity of affinity propagation is not satisfying. Thus, we propose a hybrid approach that combines affinity propagation with agglomerative hierarchical clustering to improve the clustering performance. Experiments show that compared with traditional agglomerative hierarchical clustering, the hybrid method achieves better performance on the test corpora.

show abstract

Efficient audio segmentation algorithms based on the BIC

Cited by 37 publications

References 4 publications

Speaker segmentation and clustering

Speaker segmentation and clustering

Robust Unsupervised Speaker Segmentation for Audio Diarization

Robust Speaker Clustering Using Affinity Propagation

Contact Info

Product

Resources

About