A method and device for identifying multiple organs in the abdominal cavity

Abstract

The embodiment of the invention provides an abdominal cavity multi-organ identification modeling method and device. The method comprises the steps of acquiring a medical image containing an abdominalcavity organ; pre-processing the medical image, and determining a shape distance contour surface according to the pre-processed medical image; constructing a shape distance characteristic model according to each pixel point in the medical image, the shape distance contour surface and the volume value of each abdominal cavity organ; obtaining a plurality of image blocks divided by the medical image, wherein each image block corresponds to a preset n-dimensional feature vector one by one; obtaining a first description feature and a second description feature of each image block; determining a vector value of each-dimensional feature vector according to all the feature values of all the pixel points contained in each image block, a preset first feature expression and a preset second feature expression; modeling. The device executes the above method. According to the method and the device provided by the embodiment of the invention, the abdominal cavity organ in the medical image can be accurately identified.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of medical modeling, and in particular to a method and device for identifying and modeling abdominal multiple organs. Background technique [0002] With the development of medical technology, it is particularly important to identify information in medical images (such as CT). [0003] Due to the blurred boundaries and variety of abdominal organs, and the influence of abdominal pressure, some organs (such as stomach, pancreas, liver, etc.) have large deformations. Moreover, by manually identifying abdominal organs in medical images, the efficiency is very low. The existing methods for identifying abdominal organs in medical images usually need to establish a modeling model, and then identify abdominal organs in medical images based on the modeling model. The modeling models are divided into two categories: non-a priori methods and a priori methods. The first category is systems based o...

AI Technical Summary

Problems solved by technology

Moreover, by manually identifying abdominal organs in medical images, the efficiency is very low

The existing methods for identifying abdominal organs in medical images usually need to establish a modeling model, and then identify abdominal organs in medical images based on the modeling model. The modeling models are divided into two categories: non-a priori methods and a priori methods. The first category is systems based on machine learning methods such as level set and watershed algorithm and deep learning methods. Although systems based on non-priori methods such as level set and deep learning are feasible, most of these systems are based on the underlying grayscale However, there are many types of abdominal tissues, and the gray information characteristics of each organ are not obvious. Therefore, complex parameter adjustment and post-processing steps are often required, and it is easy to cause inaccurate recognition results of the modeling model.

The other is a combined strategy method based on the statistical shape method, the probability map method, and the contextual feature classification method. However, the prior characteristics of the relative position, relative size and shape of each abdominal organ are not obvious, making modeling The model is not easy to accurately identify

Images