摘要
目的为提高医学影像检索的准确性,在PACS系统建立基于语义的医学图像检索系统。方法首先针对医学图像特征纹理丰富、灰度分辨率高的特点,结合灰度特征和纹理特征,提取出图像的混合特征;然后通过潜在主题框架将图像混合特征和高级语义相联系,实现潜在主题框架下的图像表示和检索方法。结果本系统方便医生查阅相关病例的医学图像,可有效地获取到相似的病例进行借鉴,掌握最新医疗信息。结论本系统具有实用性、可行性、扩展性,能不断提高广州市番禺区医学影像诊断水平。
Objective To improve the precision of medical image retrieval by setting up an image retrieval system in PACS. Methods In light of the rich texture feature and high gray resolution of medical image, mixing features were extracted from image combined with gray feature and texture feature, and mixing features and high-level semantics were associated in potential theme framework. Results It was convenient for doctors to consult related images, and obtain similar cases from the image retrieval system. Conclusions The image retrieval system can continuously improve diagnosis level of medical imaging of Panyu district due to its practicability, feasibility and expansibility.
出处
《临床医学工程》
2013年第7期781-782,共2页
Clinical Medicine & Engineering
基金
广州市医药卫生科技项目(201102A213058)
2011年广东省科技厅第三批科技计划项目(自筹经费-序号:7)
