Saved in:
Bibliographic Details
Main Authors: Ma, Zhiming, Gan, Shiyu, Zhao, Junhao, Li, Xianming, Pan, Qingyun, Wang, Peidong, Pan, Mingjun, Mo, Yuhao, Cheng, Jiajie, Chen, Chengxin, Cao, Zhonglun, Liu, Chonghan, Cheng, Shi
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.09915
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912708698308608
author Ma, Zhiming
Gan, Shiyu
Zhao, Junhao
Li, Xianming
Pan, Qingyun
Wang, Peidong
Pan, Mingjun
Mo, Yuhao
Cheng, Jiajie
Chen, Chengxin
Cao, Zhonglun
Liu, Chonghan
Cheng, Shi
author_facet Ma, Zhiming
Gan, Shiyu
Zhao, Junhao
Li, Xianming
Pan, Qingyun
Wang, Peidong
Pan, Mingjun
Mo, Yuhao
Cheng, Jiajie
Chen, Chengxin
Cao, Zhonglun
Liu, Chonghan
Cheng, Shi
contents Hearing-impaired individuals often face significant barriers in daily communication due to the inherent challenges of producing clear speech. To address this, we introduce the Omni-Model paradigm into assistive technology and present HI-TransPA, an instruction-driven audio-visual personal assistant. The model fuses indistinct speech with lip dynamics, enabling both translation and dialogue within a single multimodal framework. To address the distinctive pronunciation patterns of hearing-impaired speech and the limited adaptability of existing models, we develop a multimodal preprocessing and curation pipeline that detects facial landmarks, stabilizes the lip region, and quantitatively evaluates sample quality. These quality scores guide a curriculum learning strategy that first trains on clean, high-confidence samples and progressively incorporates harder cases to strengthen model robustness. Architecturally, we employs a novel unified 3D-Resampler to efficiently encode the lip dynamics, which is critical for accurate interpretation. Experiments on purpose-built HI-Dialogue dataset show that HI-TransPA achieves state-of-the-art performance in both literal accuracy and semantic fidelity. Our work establishes a foundation for applying Omni-Models to assistive communication technology, providing an end-to-end modeling framework and essential processing tools for future research.
format Preprint
id arxiv_https___arxiv_org_abs_2511_09915
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle HI-TransPA: Hearing Impairments Translation Personal Assistant
Ma, Zhiming
Gan, Shiyu
Zhao, Junhao
Li, Xianming
Pan, Qingyun
Wang, Peidong
Pan, Mingjun
Mo, Yuhao
Cheng, Jiajie
Chen, Chengxin
Cao, Zhonglun
Liu, Chonghan
Cheng, Shi
Computation and Language
Multimedia
Sound
Hearing-impaired individuals often face significant barriers in daily communication due to the inherent challenges of producing clear speech. To address this, we introduce the Omni-Model paradigm into assistive technology and present HI-TransPA, an instruction-driven audio-visual personal assistant. The model fuses indistinct speech with lip dynamics, enabling both translation and dialogue within a single multimodal framework. To address the distinctive pronunciation patterns of hearing-impaired speech and the limited adaptability of existing models, we develop a multimodal preprocessing and curation pipeline that detects facial landmarks, stabilizes the lip region, and quantitatively evaluates sample quality. These quality scores guide a curriculum learning strategy that first trains on clean, high-confidence samples and progressively incorporates harder cases to strengthen model robustness. Architecturally, we employs a novel unified 3D-Resampler to efficiently encode the lip dynamics, which is critical for accurate interpretation. Experiments on purpose-built HI-Dialogue dataset show that HI-TransPA achieves state-of-the-art performance in both literal accuracy and semantic fidelity. Our work establishes a foundation for applying Omni-Models to assistive communication technology, providing an end-to-end modeling framework and essential processing tools for future research.
title HI-TransPA: Hearing Impairments Translation Personal Assistant
topic Computation and Language
Multimedia
Sound
url https://arxiv.org/abs/2511.09915