Speech Recognition for Evaluation of phonetic Fundtion in Prosthodontic Treatment at the Dental Chair-side

语音功能是人类交际中最重要的口头功能。传统的语音评估方法需要多个评估器或特殊的设备和环境，耗时长，难以即时显示评估结果。因此，我们开发了一个新的评估系统使用语音识别，它允许定量评估。在牙科椅边说话声音的可理解性。每个声音都是用头戴式麦克风和笔记本电脑录制的。在频率和时间方向的局部变化中，突出每个声音的声时频模式，提取多个声特征平面（MAFP）模式。我们提出了一种基于213种音位和音位间特征的综合语音段的语音识别方法。该系统利用评价音节周期内正确与不正确的综合音段标签数的比值进行大周期定量音节评价。测试音被确定为音节[Shi]，它是后肺泡摩擦辅音。测试词被选为[iShimatsu]或[iShikawa]，其中包括测试音作为第二个音节。利用该系统，分析了上颌义齿内切覆盖和腭廓增加与语音功能的关系。通过增加5例受试者的切牙覆盖量，[Shi]的正确发音变得困难，并且在4 mm的覆盖量相对于已经使用的义齿，正确标签的数量明显减少。在6例全口义齿佩戴者中，下切部分腭廓线从2 mm水平增加到4 mm，表明正确标签数量显著增加。建议在假牙试制时进行言语评价，对制作无言语障碍的假牙有一定的指导意义。由于该机械系统的可理解性可以用正确标签的出现比例来表示，并且听者之间的听觉感知能力差异几乎不产生差异，因此可以说可以进行客观的、可量化的评估。少

Phonetic function is the most important oral function for human communication. Conventional speech evaluating methods require plural estimators or special equipment and environment, are time-, and it is difficult to show the results of the assessments instantly. Hence, we developed a new evaluating system using speech recognition, which allows quantitative assessment of. the intelligibility of spoken sounds at the dental chair-side.Each sound was recorded with a head-set type microphone and a laptop computer. The acoustic time-frequency pattern of each sound was accentuated in local changes in the frequency and time directions, and the pattern of multiple acoustic-feature planes (MAFP) was extracted. We proposed a speech recognition based on 213 types of integrated phonetic segments representing phoneme and inter-phoneme features. The system performed macro-period quantitative syllable evaluation using the ratio between the correct and incorrect integrated phonetic segment label number … More s during the periods of evaluated syllables.The test sound was determined as the syllable [Shi], which is the post-alveolar fricative consonant. And the test word was selected as [iShimatsu] or [iShikawa], which included the test sound as the second syllable.Using this new system, we analyzed the relationship with the incisal overjet or the augmentation of the palatal contour in a maxillary denture and the phonetic functions.By increasing the incisal overjet in 5 subjects, correct pronunciation of [Shi] became difficult, and in 4 mm overjet relative to that of the dentures already in use showed significant decrease in numbers of correct labels.In 6 complete denture wearers, the palatal contour at the incisal portion augmented horizontally from 2 to 4 mm represented significant increase in numbers of correct labels.It is suggested that speech evaluation at the time of trial arrangement in wax dentures is expected to be useful for fabricating the final denture without speech disorders.Since the intelligibility can be expressed by the appearance ratio of the correct labels in this mechanical system and almost no differences arise from variation of auditory perception ability among listeners, it can be said that objective and quantifiable assessments can be made. Less