Automatic extraction of tumor region on X-ray image of animals
Automatic extraction of tumor region on X-ray image of animals
复制标题
动物X射线图像上肿瘤区域的自动提取
DOI：
10.1109/sice.2006.315299
复制
发表时间：
2006
期刊：
影响因子：
--
通讯作者：
I. Iwayoshi
中科院分区：
文献类型：
--
作者：
D. Wakamiya;T. Tanaka;I. Kabaya;M. Kano;I. Iwayoshi
CT影像的诊断率越来越高。但由于扫描时间和扫描成本的原因，我们经常使用X光图像。X射线图像中肿瘤区域的提取比较困难，病理医生需要对大量的肿瘤图像进行诊断。因此，开发支持病理医生的CAD系统的需求越来越大。我们使用的图像是狗的图像。很多人研究人的形象，但对动物形象的研究不是很好。本文对肿瘤区域的自动提取进行了研究。我们使用了归一化相关运算。这个滤镜的模板看起来像座山。我们还使用了Quoit-Filter，检测到可能有肿瘤的区域。我们计算了肿瘤的边缘，以便于看到肿瘤。我们的方法检测到了一些肿瘤候选边缘。作为以后的工作，我们应该提取骨骼区域，并自动确定一些固定值，包括过滤器的大小
Diagnosis by CT images is increasing. But we often use X-ray images because of scanning time and scan cost. It is difficult to extract tumor region of X-ray images, and pathologists have to diagnose many images of tumor. Therefore, demand for the development of CAD system is increasing to support pathologists. Images that we use are dog images. Many people research human images, but animal images are not researched well. In this paper, automatic extraction of tumor region is studied. We used operation of normalized correlation. Template of this filter looks like mountain. We also used Quoit-filter, that detected region that had possibility of tumor. We calculated edge of tumor to see easily tumors. Our method detected some tumor candidate edge. As future work, we should extract bone region, and some fixed value, including filter size should be automatically determined