手术导航系统中可视化技术的研究

发布时间：2018-06-17 04:34

本文选题：三维可视化 + 图像分割　；参考：《天津大学》2016年硕士论文

【摘要】：手术导航系统中的可视化技术是把医学图像二维序列转换为三维实体,并根据需要为用户提供交互处理手段的技术,被广泛应用于术前规划、诊断医学及医学仿真训练等领域,具有重要的理论研究意义和实际应用价值。本文主要研究了手术导航系统中的可视化技术,提出了基于区块的多特征分割算法和基于Contourlet变换的分水岭多目标分割算法,并且完成了手术导航系统中的可视化。首先,提出了基于区块的多特征分割算法,主要包括提取图像特征、获得区块和标签传递,保证了图像特征空间的完备性;在组建区块部分,去除了低效特征,保证了算法的效率;在标签传递部分,制定了标签传递策略,并实现了基于区块的多特征图像分割算法。实验结果表明,该算法能够准确地提取目标区域。其次,提出了基于Contourlet变换的分水岭多目标分割算法,该算法对传统的分水岭分割算法进行了改进,并构造了符合人眼识别系统的对比度金字塔模型,采用基于对比度金字塔的Contourlet变换和基于纹理梯度的标记分水岭算法相结合的方式解决了过分割问题。实验结果表明,该算法能够持续稳定地产生正确的器官数目,抑制了过分割现象,并且对每个主要器官都有良好的分割效果。最后,实现了手术导航下的三维可视化系统。该系统的主要内容为:对图像进行分割以获得不同的分割区域;然后给同一区域内的像素设置相同的数值,不同区域的像素设置不同的数值,将图像标签化;最终将标签化的分割结果传递给重建模块进行重建,并获得了良好的可视化结果。
[Abstract]:The visualization technology in the surgical navigation system is to transform the medical image 2-D sequence into the three-dimensional entity, and provide the interactive processing means for the user according to the need. It has been widely used in the fields of preoperative planning, diagnostic medicine and medical simulation training and so on. It has important theoretical research significance and practical application value. In this paper, the visualization technology of surgical navigation system is studied, and a multi-feature segmentation algorithm based on blocks and a watershed multi-objective segmentation algorithm based on Contourlet transform are proposed, and the visualization in the operation navigation system is completed. First of all, a multi-feature segmentation algorithm based on blocks is proposed, which mainly includes extracting image features, obtaining blocks and label transfer, ensuring the completeness of image feature space, and removing inefficient features in building blocks. The efficiency of the algorithm is guaranteed. In the part of label transfer, the label transfer strategy is formulated, and a multi-feature image segmentation algorithm based on blocks is implemented. Experimental results show that the algorithm can accurately extract the target region. Secondly, a watershed multi-objective segmentation algorithm based on Contourlet transform is proposed. The algorithm improves the traditional watershed segmentation algorithm, and constructs a contrast pyramid model in accordance with the human eye recognition system. The problem of over-segmentation is solved by combining Contourlet transform based on contrast pyramid and marking watershed algorithm based on texture gradient. The experimental results show that the algorithm can produce the correct number of organs continuously and stably, suppress the phenomenon of over-segmentation, and have a good segmentation effect for each major organ. Finally, the three-dimensional visualization system under operation navigation is realized. The main contents of the system are as follows: segmenting the image to obtain different segmentation regions, then setting the same value for the pixels in the same region, setting different values for the pixels in different regions, and tagging the images. Finally, the tagged segmentation results are transferred to the reconstruction module, and good visualization results are obtained.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【相似文献】