面向发音运动的超声图像采集与处理的研究-软件工程专业毕业论文.docxVIP

ABSTRACT In this paper, we introduce the multi-modal articulation acquisition, including the hardware system and the software for controlling the hardware system by SDK. We use ultrasound machine, high-speed cameras, electroglottographs and microphone to acquire the articulation movement of the speaker. Because all the machines have its own sample frequency and internal clock, we design the algorithm and develop the post software to synchronize the multi-modal data precisely. Ultrasound images have serious speckle noise, which decreases the quality of ultrasound images and makes the following image-analyzing become more difficult. And we modify the standard SRAD (speckle reducing anisotropic diffusion) to improve its edge detection and suppress the intrinsic edge blurring effect of SRAD by exploiting the multidirectional separability. We experimented with both synthetic and real ultrasound images by SRAD and the proposed approach. The extracted contours in denoised images by SRAD and the proposed approach are compared in terms of the corresponding accuracy, both subjectively and objectively. The results show the proposed approach performs better than the conventional SRAD and more accurate contours can be obtained for post processing. And we optimize the iteration number by MSE and FOM as criteria. KEY WORDS：Separability, Anisotropic diffusion, Extracting contours, Multi-modal data acquisition - - 1 - - - 2 - 目 录 第一章 绪论 1 1.1 课题背景 1 1.2 研究现状 2 1.3 本文创新工作 3 1.3.1 多模态发音生理数据采集系统搭建 3 1.3.2 超声图像降噪 4 1.3.3 超声图像中舌头边界提取 5 1.4 论文结构 6 第二章 超声图像的采集 7 2.1 数据采集系统 7 2.1.1 超声采集系统 7 2.1.2 视频采集系统 10 2.1.3 喉头信号采集系统 12 2.1.4 其他设备 13 2.2 数据同步处理 13 2.2.1 超声与音频同步 13 2.2.2 视频与音频同步 14 2.2.3 喉头运动信息与音频同步 14 2.2.4 各组数据流之间的同步 14 2.3 数据演示系统 14 2.4 本章小结 15 第三章 超声图像降噪 16 3.1 超声成像技术 16 3.1.1 超声成像的发展 16 3.1.2 超声图像的噪声 16 3.2 降噪算法简述 17 3.2.1 传统的降噪算法 17 3.2.2 各项异性扩散算法 17 3.3 基于可分性的降噪算法 22 3.3.1 可分系数 22 3.3.2 多方向的可分性系数 23 3.3.3 基于多方向可分性的降噪算法 25 3.3.4 算法离散化 25 3.4 本章小结 27 第四章 实验分析 28 4.1 实验过程 28 4.1.1 实验