基于小鼠模型的器官分割方法研究

发布时间：2018-06-12 03:50

本文选题：医学图像分割 + 模糊连接度　；参考：《西安电子科技大学》2012年硕士论文

【摘要】：动物模型是现代生物医学研究中重要的实验方法和手段，分割出小动物的各个器官是各种小动物成像技术的需要，也是医学影像处理与分析的首要要求。首先，本文针对医学图像中常用的小鼠模型提出了一种结合模糊连接度和Otsu方法的交互式图像分割方法，该方法首先应用一种动态规划的方法求出种子点的模糊连接对象，然后通过Otsu方法设置阈值提取出小鼠的骨骼图像。最后以小鼠CT采集数据进行了验证，，实验结果证明了该方法的有效性。此外，针对基于高斯混合分布的医学图像分割方法，本文还提出了一种改进的各项异性平滑滤波来对图像进行预处理，然后结合直方图分析方法为分解混合密度的经典方法-期望最大算法提供良好初值，最后用期望最大算法来迭代实现对混合密度的分解。并针对小鼠的CT采集数据应用此方法提取出了其肺部图像，实验结果证明了其优越性。
[Abstract]:Animal model is an important experimental method and means in modern biomedical research. It is the need of various small animal imaging techniques to divide the various organs of small animals, and it is also the primary requirement of medical image processing and analysis.
First, an interactive image segmentation method combining fuzzy connectedness and Otsu method is proposed in this paper. Firstly, a dynamic programming method is used to find the fuzzy connection object of the seed point, and then the threshold of the mouse's skeleton image is extracted by the Otsu method. Finally, it is small. The experimental data of rat CT are verified, and the experimental results show the effectiveness of the method.
In addition, aiming at the medical image segmentation method based on Gauss mixed distribution, this paper also proposes an improved heterosexual smoothing filter to preprocess the image, and then combines the histogram analysis method to provide good initial value for the classical method of decomposing mixed density - the maximum expected maximum algorithm. Finally, the desired maximum algorithm is used to iterate the implementation. The decomposition of mixed density is carried out, and the lung image is extracted from the CT data of the mouse. The experimental results prove its superiority.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R311;TP391.41

【引证文献】