【摘要】 为在声带持续振动的情况下获取声带振动的平均速率与振幅,提出了一种基于图像特征识别的多点式信息提取方法。使用喉高速摄影设备采集图像序列后,通过确定目标振动区域和图像预处理的方法提高后续算法效率,利用声带振动特征获取声带轮廓点集,并根据向量积公式确定轮廓点集中有效点的位置,以此计算出声带振动的平均速率与振幅。实验结果表明,该方法实现了声带振动信息的量化,可为临床喉科医师诊断提供有效的参考信息。更多还原

【Abstract】 To acquire the average vibration velocity and amplitude of vocal fold information when vocal fold keeps vibrating,a method of information extraction based on image characteristics recognition is proposed.After the image sequences are extracted by throat high-speed photography,the efficiency of algorithm is improved by determining the vibration area and preprocessing images.The vocal fold contour points are collected by the feature of vocal fold vibration.Finally,the location of the effective point in the vocal fold contour points can be determined by the vector product;besides,the average velocity and the amplitude of vocal fold can be calculated.According to the results of the experiment,the quantification of vocal fold vibration information is realized by this method.It can also provide some effective reference information to clinical otolaryngology doctors.更多还原