基于胸片局部图像配准的身份识别技术研究

Identity Recognition Based on Local Registration of Chest Radiography

在医院的PACS系统中,胸片图像归档错误医生可通过对比新加入PACS的图像和已有的图像来核查图像归档错误。探索使用图像处理与分析的技术实现这类问题的计算机自动处理。介绍了胸片中解剖组织的粗略定位方法,使用图像配准方法对图像局部相似性进行了比较,并结合了人工神经网络技术综合局部成果,得到良好的整体识别效果.

In a PACS system, doctors can avoid the mistakes in filing chest radiographs by comparing the new image with the old ones from the same patient. An automatic method with digital image processing technology is introduced in the paper, which is realized by general location of anatomical structures, local image registration and BP artificial neural network, so as to achieve good results for identity recognition.