The paper proposes an approach for extending deep neural networks-based solutions to closed-set speaker identification toward the open-set problem. The idea is built on the characteristics of deep neural networks trained for the classification tasks, where there is a layer consisting of a set of deep features extracted from the analyzed inputs. By extracting this vector and performing anomaly detection against the set of known speakers, new speakers can be detected and modeled for further re-identification. The approach is tested on the basis of NeMo toolkit with SpeakerNet architecture. The algorithm is shown to be working with multiple new speakers introduced.
Authors
Additional information
- DOI
- Digital Object Identifier link open in new tab 10.1007/978-3-031-16159-9_14
- Category
- Publikacja monograficzna
- Type
- rozdział, artykuł w książce - dziele zbiorowym /podręczniku w języku o zasięgu międzynarodowym
- Language
- angielski
- Publication year
- 2022
