Collection:
In this paper, a task-driven multilevel framework (TDMF) is proposed for end-to-end speaker verification. The TDMF has four layers, and each layer has different effects on speaker models or representations to implement the functions of universal backgroun
- IEEE MemberUS $11.00
- Society MemberUS $0.00
- IEEE Student MemberUS $11.00
- Non-IEEE MemberUS $15.00
Videos in this product
Tdmf: Task-Driven Multilevel Framework For End-To-End Speaker Verification
In this paper, a task-driven multilevel framework (TDMF) is proposed for end-to-end speaker verification. The TDMF has four layers, and each layer has different effects on speaker models or representations to implement the functions of universal backgroun