VocaSense-X框架：通过智能诊断与康复推进嗓音医学

文献信息

中文摘要

嗓音障碍是一个重要的临床问题，影响全球数百万人，并可能损害其生活质量、职业功能和沟通能力。当前的诊断方法可能基于主观临床检查和分散的康复方案，可能缺乏准确性和可扩展性。本研究介绍了VocaSense-X框架，这是一个通过整合诊断与康复来改善嗓音医学的综合智能系统。该框架使用多模态信号处理、深度学习模型和自适应反馈机制，对嗓音病理（包括发声困难、声带病变和神经源性嗓音障碍）进行准确、数据驱动的分析。VocaSense-X是一个将实时声学特征提取流程与混合卷积-循环神经网络相结合的模型，在广泛的患者群体中实现了优越的分类性能。该框架包括一个个性化康复模块，该模块根据对嗓音的持续监测动态调整治疗性练习，从而能够跟踪患者随时间推移的进展并支持诊断。使用基准嗓音障碍数据集进行的实验分析表明，在相同的实验设置下，VocaSense-X的诊断性能显著优于所测试的任何其他比较架构，准确率>94%。此外，临床医生可用性测试证实了该框架在临床和远程治疗环境中的可用性。本文提出的系统旨在填补智能计算与临床嗓音医学之间的空白，具有可扩展性和循证性，可用于早期检测和结构化嗓音康复。VocaSense-X对耳鼻喉科、言语-语言病理学和数字健康平台的意义使其成为迈向AI驱动嗓音医疗创新的关键一步。

英文摘要

Voice disorders are a significant clinical issue that affect millions of people worldwide and can impair their quality of life, professional functioning, and communication. Current methods of diagnosis may be based on subjective clinical examination and disjointed rehabilitation regimens, which may lack accuracy and scalability. The study introduces the VocaSense-X Framework, a comprehensive intelligent system for improving voice medicine through integrated diagnosis and rehabilitation. The framework uses multimodal signal processing, deep learning models, and adaptive feedback mechanisms to provide accurate, data-driven analysis of vocal pathologies, including dysphonia, vocal fold lesions, and neurogenic voice disorders. VocaSense-X is a model that combines a real-time acoustic feature extraction pipeline with a hybrid convolutional-recurrent neural network, achieving superior classification performance across a wide range of patient populations. The framework includes a personalized rehabilitation module that dynamically adjusts therapeutic exercises based on continuous monitoring of the voice, enabling tracking of the patient's progress over time and supporting diagnosis. The experimental analysis using benchmark voice disorder datasets demonstrates that the diagnostic performance of VocaSense-X is significantly better than that of any other comparator architecture tested under the same experimental setup, with an accuracy of > 94. Furthermore, clinician usability tests confirm the framework's usability in both clinical and teletherapy settings. The system proposed herein is designed to fill the gap between intelligent computation and clinical voice medicine, is scalable and evidence-based, and can be used for early detection and structured voice rehabilitation. The implications of VocaSense-X for otolaryngology, speech-language pathology, and digital health platforms make it a pivotal step towards AI-driven voice healthcare innovations.