摘要
目的研究利用计算机处理分析DICOM格式CT图像的方法及软件实现,探讨肺结节的计算机辅助检测方法。方法利用VC++结合OpenGL编程读取DICOM格式的CT图像,然后运用多种数字图像处理方法在脱离影像工作站的情况下处理分析图像,并采用多尺度增强和交互式区域增长结合神经网络的方法实现肺结节的计算机辅助检测。结果本研究提出并实现了一个处理分析CT图像并对肺结节进行计算机辅助检测的算法系统,并应用于影像诊断和图像处理教学,实际处理了含40个结节的图像40幅,检测出结节38个,有效率95%。结论本研究利用VC++结合OpenGL编程实现了CT图像的计算机处理分析和肺结节的计算机辅助检测,程序界面友好,操作方便,适合教学使用。
Objective： To research the methods of processing and analyzing DICOM CT image, probe into computeraided detection of lung nodules and the software realization. Methods： Firstly, the DICOM image file were opened by programming with VC＋＋ and OpenGL, then a series of image processing algorithm on the image were performed. Lastly, multi-scale enhance filters and interactive region-growing algorithm followed by an BP ANN classifier were used to enhance, segment the suspicious nodule region and reduce false positives. Results： This paper proposed and realized a processing and analysing system of DICOM lung CT image on the basis of summarizing previous research. 40 CT images that contained 40 nodules were used to test the system, and 95% （38/40） nodules were detected correctly. The system now has been used in the teaching and researching of medical image diagnosis and image processing in our department. Conclusion The system proposed by this paper proves to be practical.
出处
《泰山医学院学报》
CAS
2005年第3期200-202,共3页
Journal of Taishan Medical College
