Speaker Diarization

Speaker Diarization is the task of segmenting and co-indexing audio recordings by speaker. The way the task is commonly defined, the goal is not to identify known speakers, but to co-index segments that are attributed to the same speaker; in other words, diarization implies finding speaker boundaries and grouping segments that belong to the same speaker, and, as a by-product, determining the number of distinct speakers. In combination with speech recognition, diarization enables speaker-attributed speech-to-text transcription.

Source: Improving Diarization Robustness using Diversification, Randomization and the DOVER Algorithm

Papers

Recently Added Most Hyped Most Active Needs Verification Most Verified

Showing 61–70 of 328 papers

Title	Date	Tasks	Status	Score
CountNet: Estimating the Number of Concurrent Speakers Using Supervised Learning Speaker Count Estimation	Oct 28, 2018	blind source separationspeaker-diarization	CodeCode Available	5
Self-Supervised Metric Learning With Graph Clustering For Speaker Diarization	Sep 14, 2021	ClusteringGraph Clustering	CodeCode Available	5
Probabilistic embeddings for speaker diarization	Apr 6, 2020	Clusteringspeaker-diarization	CodeCode Available	5
Powerset multi-class cross entropy loss for neural speaker diarization	Oct 19, 2023	Multi-class ClassificationMulti-Label Classification	CodeCode Available	5
A Comprehensive Evaluation of Incremental Speech Recognition and Diarization for Conversational AI	Dec 1, 2020	Automatic Speech RecognitionAutomatic Speech Recognition (ASR)	CodeCode Available	5
Compositional embedding models for speaker identification and diarization with simultaneous speech from 2+ speakers	Oct 22, 2020	speaker-diarizationSpeaker Diarization	CodeCode Available	5
Sortformer: Seamless Integration of Speaker Diarization and ASR by Bridging Timestamps and Tokens	Sep 10, 2024	speaker-diarizationSpeaker Diarization	CodeCode Available	5
Compositional Clustering: Applications to Multi-Label Object Recognition and Speaker Identification	Sep 9, 2021	ClusteringFew-Shot Learning	CodeCode Available	5
End-to-End Supervised Hierarchical Graph Clustering for Speaker Diarization	Jan 23, 2024	ClusteringGraph Clustering	CodeCode Available	5
On Out-of-Distribution Detection for Audio with Deep Nearest Neighbors	Oct 27, 2022	Automatic Speech RecognitionAutomatic Speech Recognition (ASR)	CodeCode Available	5

Show:10 25 50

← PrevPage 7 of 33Next →

All datasets CALLHOME NIST-SRE 2000 AMI Lapel AMI MixHeadset CH109 DIHARD ETAPE AMI CALLHOME-109 AliMeeting DIHARD II Hub5'00 CallHome

Benchmark Results

#	Model	Metric	Claimed	Verified	Status
1	COS+NJW-SC (Oracle SAD)	DER(%)	24.05	—	Unverified
2	EEND	DER(%)	23.07	—	Unverified
3	COS+AHC (Oracle SAD)	DER(%)	21.13	—	Unverified
4	SA-EEND (2-spk, no-adapt)	DER(%)	12.66	—	Unverified
5	EEND-OLA	DER(%)	12.57	—	Unverified
6	SA-EEND (2-spk, adapted)	DER(%)	10.76	—	Unverified
7	TOLD	DER(%)	10.14	—	Unverified
8	COS+B-SC (Oracle SAD)	DER(ig olp)	8.78	—	Unverified
9	PLDA+AHC (Oracle SAD)	DER(ig olp)	8.39	—	Unverified
10	COS+NME-SC (Oracle SAD)	DER(ig olp)	7.29	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	x-vector (PLDA + AHC)	DER(%)	8.39	—	Unverified
2	TitaNet-L (NME-SC)	DER(%)	6.73	—	Unverified
3	TitaNet-M (NME-SC)	DER(%)	6.47	—	Unverified
4	TitaNet-S (NME-SC)	DER(%)	6.37	—	Unverified
5	x-vector (MCGAN)	DER(%)	5.73	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	ECAPA (SC)	DER(%)	2.36	—	Unverified
2	TitaNet-L (NME-SC)	DER(%)	2.03	—	Unverified
3	TitaNet-S (NME-SC)	DER(%)	2	—	Unverified
4	TitaNet-M (NME-SC)	DER(%)	1.99	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	TitaNet-S (NME-SC)	DER(%)	2.22	—	Unverified
2	TitaNet-M (NME-SC)	DER(%)	1.79	—	Unverified
3	ECAPA (SC)	DER(%)	1.78	—	Unverified
4	TitaNet-L (NME-SC)	DER(%)	1.73	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	x-vector (PLDA + AHC)	DER(%)	9.72	—	Unverified
2	TitaNet-L (NME-SC)	DER(%)	1.19	—	Unverified
3	TitaNet-M (NME-SC)	DER(%)	1.13	—	Unverified
4	TitaNet-S (NME-SC)	DER(%)	1.11	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	Baseline (the best result in the literature as of Oct.2019)	DER(%)	11.2	—	Unverified
2	pyannote (MFCC)	DER(%)	10.5	—	Unverified
3	pyannote (waveform)	DER(%)	9.9	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	Baseline	DER(%)	7.7	—	Unverified
2	pyannote (MFCC)	DER(%)	5.6	—	Unverified
3	pyannote (waveform)	DER(%)	4.9	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	pyannote (MFCC)	DER(%)	6.3	—	Unverified
2	pyannote (waveform)	DER(%)	6	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	d-vector + spectral	DER(%)	12.54	—	Unverified
2	titanet-s	DER(%)	1.11	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	SOND	DER(%)	4.46	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	UIS-RNN-SML	DER(%)	27.3	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	UIS-RNN	V	10.6	—	Unverified