EvoTSE improves target speaker extraction by continuously evolving the enrollment via reliability-filtered retrieval of high-confidence historical estimates, reducing speaker confusion and boosting out-of-domain performance without extra annotated data.
A re- gion based non-overlapping reference speech estimation me thod for speaker extraction,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
EvoTSE: Evolving Enrollment for Target Speaker Extraction
EvoTSE improves target speaker extraction by continuously evolving the enrollment via reliability-filtered retrieval of high-confidence historical estimates, reducing speaker confusion and boosting out-of-domain performance without extra annotated data.