Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2602.08696 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918344369635328 |
|---|---|
| author | Wang, Haoshen Zhong, Xueli Lin, Bingbing Huang, Jia Pan, Xingduo Liang, Shengxiang Wang, Nizhuan Siok, Wai Ting |
| author_facet | Wang, Haoshen Zhong, Xueli Lin, Bingbing Huang, Jia Pan, Xingduo Liang, Shengxiang Wang, Nizhuan Siok, Wai Ting |
| contents | Dysarthric speech exhibits high variability and limited labeled data, posing major challenges for both automatic speech recognition (ASR) and assistive speech technologies. Existing approaches rely on synthetic data augmentation or speech reconstruction, yet often entangle speaker identity with pathological articulation, limiting controllability and robustness.
In this paper, we propose ProtoDisent-TTS, a prototype-based disentanglement TTS framework built on a pre-trained text-to-speech backbone that factorizes speaker timbre and dysarthric articulation within a unified latent space. A pathology prototype codebook provides interpretable and controllable representations of healthy and dysarthric speech patterns, while a dual-classifier objective with a gradient reversal layer enforces invariance of speaker embeddings to pathological attributes. Experiments on the TORGO dataset demonstrate that this design enables bidirectional transformation between healthy and dysarthric speech, leading to consistent ASR performance gains and robust, speaker-aware speech reconstruction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_08696 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Prototype-Based Disentanglement for Controllable Dysarthric Speech Synthesis Wang, Haoshen Zhong, Xueli Lin, Bingbing Huang, Jia Pan, Xingduo Liang, Shengxiang Wang, Nizhuan Siok, Wai Ting Sound Computation and Language Dysarthric speech exhibits high variability and limited labeled data, posing major challenges for both automatic speech recognition (ASR) and assistive speech technologies. Existing approaches rely on synthetic data augmentation or speech reconstruction, yet often entangle speaker identity with pathological articulation, limiting controllability and robustness. In this paper, we propose ProtoDisent-TTS, a prototype-based disentanglement TTS framework built on a pre-trained text-to-speech backbone that factorizes speaker timbre and dysarthric articulation within a unified latent space. A pathology prototype codebook provides interpretable and controllable representations of healthy and dysarthric speech patterns, while a dual-classifier objective with a gradient reversal layer enforces invariance of speaker embeddings to pathological attributes. Experiments on the TORGO dataset demonstrate that this design enables bidirectional transformation between healthy and dysarthric speech, leading to consistent ASR performance gains and robust, speaker-aware speech reconstruction. |
| title | Prototype-Based Disentanglement for Controllable Dysarthric Speech Synthesis |
| topic | Sound Computation and Language |
| url | https://arxiv.org/abs/2602.08696 |