Recognition method and segmentation method of organs in medical images

Abstract

The invention discloses a method for identifying organs in a medical image, comprising: acquiring a medical image to be processed, splitting the medical image into several two-dimensional images in the directions of X, Y, and Z axes, and Set the size of the detection window; use the detection window to traverse and detect the two-dimensional image according to the set detection step, and obtain the detection results in the X, Y and Z axis directions; perform result fusion on the detection results , keep the pixel points that are detected as positive in the three directions of the X, Y and Z axes, so as to determine the boundary of the target organ. The method for identifying organs in medical images of the invention can accurately and quickly identify areas where target organs are evenly located, determine the boundaries of target organs, and has strong self-adaptive ability. In addition, the present invention also provides a method for segmenting organs in medical images.

Description

【Technical field】 [0001] The invention relates to the field of medical image processing, in particular to an identification method and a segmentation method of organs in medical images. 【Background technique】 [0002] With the increasing maturity of medical imaging technology and the wide application of various medical imaging equipment in hospitals, information images of human internal tissues can be obtained conveniently and non-destructively, and these information can be effectively processed by image processing technology to assist doctors in diagnosis. Even surgical planning, etc., has significant social benefits and broad application prospects. For example, a computed tomography (CT) image is a matrix composed of a certain number of pixels arranged in different gray levels from black to white, and the pixels can reflect the X-ray absorption coefficient of the corresponding voxel, while different gray levels reflect the organs. Or the degree to which tissues absorb X-r...

AI Technical Summary

Problems solved by technology

[0003] Manually identifying and marking the body part where the current CT image is located by the doctor requires a lot of repetitive work by the doctor, and the efficiency is low

The existing methods for automatic recognition of body parts in CT images can be mainly divided into three types: (1) body part recognition based on the header information of Digital Imaging and Communications in Medicine (DICOM) files [1], usually DICOM The file header contains the label information of the CT image scan, but due to differences in various cultures and languages, labels recorded in different languages ​​will increase the difficulty of accurately identifying the DICOM file header information, and even wrong DICOM file header information will lead to identification errors; (2) The method based on gray value features is mainly based on the different attenuation degrees of X-rays when passing through different tissue components of the body. The gray value distribution of different tissue components in the CT image is different. The method based on gray value features is based on the body The prior knowledge of the gray value distribution of the main tissue components in the CT image can be used to divide the body parts. However, this method has a low recognition rate for the head and pelvis [2]; (3) Based on the machine learning method, the method It is mainly divided into two stages of training and testing. In the training stage, the Haar image features corresponding to the key organs of the body are extracted, and a large number of positive and negative samples are constructed. By training the AdaBoost classifier, the effective Haar feature sequences corresponding to the organs and their corresponding weights are extracted. In the test phase, input the image to be tested, calculate the Haar feature value of the image, compare it with the existing training results, and determine whether the image is a positive sample [3]. This method needs to extract the Haar feature of the image. The window performs multiple upsampling or downsampling, and the number of Haar features used is large, which has the problems of repeated calculation, large amount of calculation, and low calculation efficiency.

Images