SoCov pooling compresses a covariance matrix into a vector via a semi-orthogonal parametric layer and reports large EER gains on SRE21, despite an incorrect derivation.
Curelaru, ”Front-End Factor Analysis For Speaker Verification,” in 2018 International Conference on Communications (COMM), 2018, pp
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.AS 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
SoCov: Semi-Orthogonal Parametric Pooling of Covariance Matrix for Speaker Recognition
SoCov pooling compresses a covariance matrix into a vector via a semi-orthogonal parametric layer and reports large EER gains on SRE21, despite an incorrect derivation.