Automatic classification method and device for CT slice images

The invention discloses an automatic classification method and device for CT slice images. The method includes: extracting a body contour square frame from the CT slice image; then, dividing the bodycontour square frame into different sub square frames; integrating the gradients and the LBP feature statistical information in different sub-frames to generate global feature vectors corresponding tothe CT slice images; comparing and calculating with a template set to obtain a final classification result; therefore, classification of CT slice images is rapidly achieved with a small calculation amount and low calculation complexity, the requirement for real-time processing can be met, the requirement for performance of software and hardware is lowered, cost can be saved, development difficulty is lowered, and the requirement for a high-speed large-scale data processing mode is met. 本发明公开了一种CT切片图像的自动分类方法和装置，从CT切片图像之中提取身体轮廓方框，再分割为不同的子方框，整合不同子方框内的梯度和LBP特征统计信息生成对应CT切片图像的全局特征向量，通过与模板集进行对比计算从而获得最终的分类结果，从而实现以较小的计算量，较低的计算