Search
Abstract
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.
Citations
-
0
CrossRef
-
0
Web of Science
-
0
Scopus
Cite as
Full text
download paper
downloaded 58 times
- Publication version
- Accepted or Published Version
- License
- Copyright (2023 The Author(s), under exclusive license to Springer Nature Switzerland AG)
Keywords
Details
- Category:
- Monographic publication
- Type:
- rozdział, artykuł w książce - dziele zbiorowym /podręczniku w języku o zasięgu międzynarodowym
- Language:
- English
- Publication year:
- 2022
- Bibliographic description:
- Affek M., Tatara M.: Open-Set Speaker Identification Using Closed-Set Pretrained Embeddings// Intelligent and Safe Computer Systems in Control and Diagnostics/ : , 2022, s.167-177
- DOI:
- Digital Object Identifier (open in new tab) 10.1007/978-3-031-16159-9_14
- Sources of funding:
-
- Free publication
- Verified by:
- Gdańsk University of Technology
seen 155 times