Auto-contouring of the rectum on megavoltage computed tomography images
Auto-contouring of the rectum on megavoltage computed tomography images
复制标题
兆伏计算机断层扫描图像上直肠的自动轮廓
DOI：
--
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
M. Romanchikova
中科院分区：
文献类型：
--
作者：
M. Sutcliffe;K. Harrison;J. E. Scaife;M. A. Parker;M. Romanchikova
本报告描述了作为VoxTox程序的一部分开发的自动轮廓算法的详细信息[1]。其目的是自动勾画直肠轮廓，从而了解前列腺癌放射治疗对直肠的毒性。轮廓绘制的基础是Chan-Vese算法[2]，该算法在MatLab编码环境中以2D形式实现[3]。给出了起始轮廓、图像窗口和不良轮廓识别的详细信息。该方法是在训练数据集上开发的，并在测试集上进行了评估。结果显示前景看好。已在VoxTox计划内对该方法的准确性进行了进一步的量化，并将单独报告。在巨容量CT图像上自动绘制直肠轮廓
This report describes details of an auto-contouring algorithm developed as part of the VoxTox programme [1]. The aim is to contour the rectum automatically, so as to understand toxicity to the rectum associated with radiotherapy treatment of prostate cancer. The basis of the contouring is the Chan-Vese algorithm [2], which is implemented in 2D within the Matlab coding environment [3]. Details are given of the start contour, image windowing, and identification of poor contours. The method has been developed on a training set of data, and evaluated on a test set. Results show promise. Further quantification of the accuracy of the method has been undertaken within the VoxTox program, and will be reported separately. Auto-contouring of the rectum on megavoltage computed tomography images