Guardado en:
Detalles Bibliográficos
Autores principales: Chen, Zhiyong, Wu, Shuhang, Duan, Yingjie, Xu, Xinkang, Hu, Xinhui
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:https://arxiv.org/abs/2604.13605
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917409707786240
author Chen, Zhiyong
Wu, Shuhang
Duan, Yingjie
Xu, Xinkang
Hu, Xinhui
author_facet Chen, Zhiyong
Wu, Shuhang
Duan, Yingjie
Xu, Xinkang
Hu, Xinhui
contents This paper proposes an improved approach for open-set speaker identification based on pretrained speaker foundation models. Building upon the previous Speaker Reciprocal Points Learning framework (V1), we first introduce an enhanced open-set learning objective by integrating reciprocal points learning with logit normalization (LogitNorm) and incorporating adaptive anchor learning to better constrain target speaker representations and improve robustness. Second, we propose a model fusion strategy to stabilize and enhance the few-shot tuning process, effectively reducing result randomness and improving generalization. Furthermore, we introduce a model selection method to ensure optimal performance in model fusion. Experimental evaluations on the VoxCeleb, ESD and 3D-Speaker datasets demonstrate the effectiveness and robustness of the proposed method under diverse conditions. On a newly proposed Vox1-O-like test set, our method reduces the EER from 1.28% to 0.09%, achieving a relative reduction of approximately 93%.
format Preprint
id arxiv_https___arxiv_org_abs_2604_13605
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SpeakerRPL v2: Robust Open-set Speaker Identification through Enhanced Few-shot Foundation Tuning and Model Fusion
Chen, Zhiyong
Wu, Shuhang
Duan, Yingjie
Xu, Xinkang
Hu, Xinhui
Audio and Speech Processing
This paper proposes an improved approach for open-set speaker identification based on pretrained speaker foundation models. Building upon the previous Speaker Reciprocal Points Learning framework (V1), we first introduce an enhanced open-set learning objective by integrating reciprocal points learning with logit normalization (LogitNorm) and incorporating adaptive anchor learning to better constrain target speaker representations and improve robustness. Second, we propose a model fusion strategy to stabilize and enhance the few-shot tuning process, effectively reducing result randomness and improving generalization. Furthermore, we introduce a model selection method to ensure optimal performance in model fusion. Experimental evaluations on the VoxCeleb, ESD and 3D-Speaker datasets demonstrate the effectiveness and robustness of the proposed method under diverse conditions. On a newly proposed Vox1-O-like test set, our method reduces the EER from 1.28% to 0.09%, achieving a relative reduction of approximately 93%.
title SpeakerRPL v2: Robust Open-set Speaker Identification through Enhanced Few-shot Foundation Tuning and Model Fusion
topic Audio and Speech Processing
url https://arxiv.org/abs/2604.13605