Renal cortex locating method based on statistical shape model

Abstract

The invention discloses a renal cortex locating method based on a statistical shape model. The method includes a training stage and a renal cortex locating stage. The method is characterized in that in the training stage, artificial marks L1 and L2 of a kidney of each three-dimensional CT image in a training data set are made, binary data in L1 and L2 marking regions is converted to surface data M1 and M2 in a corresponding manner, interior surface data of renal cortex is calculated, and a renal cortex statistical shape model is established. According to the method, in the training stage, the renal cortex statistical shape model is established to perform statistics on the change mode of renal cortex, and the accuracy and the rapidity of locating of renal cortex are improved by employing an iteration model deformation algorithm.

Description

technical field [0001] The invention belongs to the field of medical imaging algorithms, and in particular relates to a renal cortex positioning method based on a statistical shape model. Background technique [0002] The renal cortex is an important part of the kidney. About 1.9% of adults are diagnosed with visceral diseases, and a considerable number of them die from nephritis, nephrotic syndrome, nephropathy and other diseases related to the renal cortex. Therefore The diagnostic research on the renal cortex is of great significance in the study of renal diseases. [0003] The existing medical imaging technology can not do a good job of identifying and positioning the renal cortex. The main reason is that the kidney is different from other organs such as the liver, and the anatomical structure of the kidney is relatively complicated. The kidney has four anatomical structures, among which the renal cortex and renal column are connected, and the renal cortex and renal col...

AI Technical Summary

Problems solved by technology

[0003] The existing medical imaging technology can not do a good job of identifying and positioning the renal cortex. The main reasons are: The kidney is different from other organs such as the liver, and the anatomical structure of the kidney is relatively complex

The kidney has four anatomical structures, among which the renal cortex and renal column are connected, and the renal cortex and renal column have similar light reflection intensity, so it is difficult to distinguish in medical imaging

In addition, the kidney and adjacent organs, such as the liver and spleen, often overlap in the image, so it is difficult to accurately model in medical imaging

The existing localization methods based on learning algorithms cannot accurately locate the renal cortex, and the model training efficiency is not high

Images