基于CT图像的肺结节计算机辅助检测技术的研究进展

Progress in Computer-Aided Detection for Pulmonary Nodule Using CT Image

目的:通过对国内外在基于CT图像的肺结节计算机辅助检测技术(Computer-Aided Detection,CAD)领域的研究状况及其研究进展的综述,以期能对CAD研究提供借鉴。方法:根据对近年来国内外报道的CAD文献的分析和深入研究,结合作者在CAD方面的研究体会,本文对CAD国内外研究进展情况、存在的问题及可能的解决方案进行了详细的论述。结果:通过对目前国内外CAD研究中存在的问题及原因的深入分析,我们认为在建立能够描述不同类型肺结节的数学模型的基础上,基于大样本病历的数据库,设计有效的检测方法是提高肺结节检测效率的关键。结论:CAD能有效辅助放射科医生从CT图像中检测出肺结节,从而为肺癌的早期诊断奠定基础,但由于CAD研究尚存在的局限性,目前报道的CAD研究离临床的实际需要尚有较大的差距。

Objective： This paper summarized the research status and progress of computer-aided detection （CAD） for pulmonary nodule using CT image in home and abroad, and we hope this can be a useful reference for CAD researchers. Method： According to the deep analysis and research for recent reports and literatures about CAD and combined with auctorial research experience for CAD, the research progress about CAD in home and abroad, the existed problems in CAD research and the possible resolve scheme for these problems were discussed in detail in this paper. Result： Through the deep analysis for the problems existed in CAD research, we think that the key to enhance the lung nodule detection efficiency is to design the effective detection method, which was based on the large sample database and the fit model that can correct depict the nodule for different types. Conclusion： CAD can help radiologist to find pulmonary nodule efficiently, which is the foundation for eraly diagnosis of lung cancer. Owing to the CAD research still had the localization, the gap between the CAD research reported at present and the clinical practical demand is still big enough.