EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

HGRN2-Based Personal Voice Activity Detection: A Lightweight Recurrent Framework for Inference and Training

Tzu-Wei Wang, Tai-You Chen, Chien-Chia Chiu, Berlin Chen, Jeih-Weih Hung

Electronics · 08.04.2026 · 15 (8) · 1561

DOI: 10.3390/electronics15081561 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronics, online erschienen am 08.04.2026. Verfasst von Tzu-Wei Wang und 4 weiteren Autor:innen. Die Arbeit nennt 23 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study presents HGRN2-based Flexible Dynamic Encoder Personal VAD (FDE-HGRN2), a recurrent framework for personal voice activity detection (PVAD). Building on the original LSTM-based FDE-RNN backbone, we replace all recurrent modules with the recently introduced HGRN2 gated linear RNN and adopt a cosine-annealing learning rate schedule to improve both detection accuracy and efficiency. HGRN2 uses gated linear recurrence with non-parametric state expansion, enlarging the recurrent state without increasing the number of trainable parameters and enabling more expressive long-range temporal modeling than conventional LSTMs. We evaluate FDE-HGRN2 on a LibriSpeech-derived PVAD benchmark, where multi-speaker mixtures are constructed by concatenating one to three speakers per utterance and randomly designating a target speaker, following established PVAD data construction practices to ensure direct comparability with prior work. The system uses 40-dimensional Mel-filterbank features as acoustic inputs and conditions the detector on 256-dimensional d-vector embeddings extracted from a pretrained speaker verification network. Experimental results show that FDE-HGRN2 consistently outperforms the original FDE-RNN baseline and several state-of-the-art PVAD models in terms of mean Average Precision and frame-level accuracy, while reducing the parameter count of the recurrent backbone by roughly 15% and yielding substantially smaller models than many competing systems. These findings indicate that HGRN2 provides a more temporally expressive and parameter-efficient alternative to LSTM for PVAD, offering a favorable accuracy–efficiency trade-off for real-world, deployment-oriented personalized speech interfaces.

Zitieren

APA

Wang, T., Chen, T., Chiu, C., Chen, B., & Hung, J. (2026). HGRN2-Based Personal Voice Activity Detection: A Lightweight Recurrent Framework for Inference and Training. Electronics, 15(8), 1561. https://doi.org/10.3390/electronics15081561

Vancouver

Wang T, Chen T, Chiu C, Chen B, Hung J. HGRN2-Based Personal Voice Activity Detection: A Lightweight Recurrent Framework for Inference and Training. Electronics. 2026;15(8):1561. doi:10.3390/electronics15081561

BibTeX
@article{Wang2026,
  author = {Wang, Tzu-Wei and Chen, Tai-You and Chiu, Chien-Chia and Chen, Berlin and Hung, Jeih-Weih},
  title = {HGRN2-Based Personal Voice Activity Detection: A Lightweight Recurrent Framework for Inference and Training},
  journal = {Electronics},
  year = {2026},
  volume = {15},
  number = {8},
  pages = {1561},
  doi = {10.3390/electronics15081561}
}
Kostenlos starten →