Menu
Publications
2026
2025
2024
2023
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
Editor-in-Chief
Nikiforov
Vladimir O.
D.Sc., Prof.
Partners
doi: 10.17586/2226-1494-2026-26-5-1077-1079
Cascade method for voice-based speaker search using cross-model comparisons
Read the full article
Article in Russian
For citation:
Abstract
For citation:
Khmelev N.A. Cascade method for voice-based speaker search using cross-model comparisons. Scientific and Technical Journal of Information Technologies, Mechanics and Optics, 2026, vol. 26, no. 5, pp. 1077–1079 (in Russian). doi: 10.17586/2226-1494-2026-26-5-1077-1079
Abstract
We address voice-based speaker search in a large gallery of voices. Modern models pretrained with self-supervised learning, such as wav2vec 2.0, provide high search accuracy but require substantial computation for extracting speaker embeddings. To reduce this cost, we propose a cascaded method: the gallery is enrolled using a fast student model, dWhisper-ASV, while queries and candidate verification are performed by a teacher model, wav2vec 2.0 XLSR-1B. The alignment of embedding spaces enables cross-model cosine similarity. On ClovaCall, Speech-MASSIVE, and a proprietary Russian corpus, RusIVR_test, the cascade nearly matches teacher-only F1-score: 77.21 % vs. 77.24 %, 83.05 % vs. 83.16 %, and 92.26 % vs. 92.63 %. Only 1.5 %, 0.2 %, and 0.11 % of records reach the second stage and the enrollment speeds up eightfold.
Keywords: speaker search, cascade method, cross-model comparison, speaker embeddings, gallery enrollment

