Spatially-Augmented Sequence-to-Sequence Neural Diarization for Meetings
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910044517302272 |
|---|---|
| author | Li, Li Cheng, Ming Liu, Juan Li, Ming |
| author_facet | Li, Li Cheng, Ming Liu, Juan Li, Ming |
| contents | This paper proposes a Spatially-Augmented Sequence-to-Sequence Neural Diarization (SA-S2SND) framework, which integrates direction-of-arrival (DOA) cues estimated by SRP-DNN into the S2SND backbone. A two-stage training strategy is adopted: the model is first trained with single-channel audio and DOA features, and then further optimized with multi-channel inputs under DOA guidance. In addition, a simulated DOA generation scheme is introduced to alleviate dependence on matched multi-channel corpora. On the AliMeeting dataset, SA-S2SND consistently outperform the S2SND baseline, achieving a 7.4% relative DER reduction in the offline mode and over 19% improvement when combined with channel attention. These results demonstrate that spatial cues are highly complementary to cross-channel modeling, yielding good performance in both online and offline settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_09505 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Spatially-Augmented Sequence-to-Sequence Neural Diarization for Meetings Li, Li Cheng, Ming Liu, Juan Li, Ming Audio and Speech Processing This paper proposes a Spatially-Augmented Sequence-to-Sequence Neural Diarization (SA-S2SND) framework, which integrates direction-of-arrival (DOA) cues estimated by SRP-DNN into the S2SND backbone. A two-stage training strategy is adopted: the model is first trained with single-channel audio and DOA features, and then further optimized with multi-channel inputs under DOA guidance. In addition, a simulated DOA generation scheme is introduced to alleviate dependence on matched multi-channel corpora. On the AliMeeting dataset, SA-S2SND consistently outperform the S2SND baseline, achieving a 7.4% relative DER reduction in the offline mode and over 19% improvement when combined with channel attention. These results demonstrate that spatial cues are highly complementary to cross-channel modeling, yielding good performance in both online and offline settings. |
| title | Spatially-Augmented Sequence-to-Sequence Neural Diarization for Meetings |
| topic | Audio and Speech Processing |
| url | https://arxiv.org/abs/2510.09505 |