CT image bone condition detection method and device based on convolutional neural network

Abstract

The invention belongs to the technical field of medical image recognition, and particularly relates to a CT image bone condition detection method and device based on a convolutional neural network. The method comprises the steps of designing a U-Net convolutional neural network model used for segmenting a lumbar vertebra part in an original CT image to obtain a marked image, and designing a DenseNet convolutional neural network model used for performing bone condition classification on the segmented image; training the two convolutional neural networks by using a clinical training data set; segmenting the original CT image by using the trained U-Net convolutional neural network model to obtain a corresponding marked image, and cutting and zooming the original CT image and the marked imageto obtain a segmented image; and classifying the segmented images by using the trained DenseNet convolutional neural network model to obtain bone condition information corresponding to the original CTimage. Dependence on auxiliary hardware and a special bone detection process is reduced, bone condition detection can be rapidly and conveniently achieved, and the bone mass loss diagnosis effect inthe clinical environment is improved.

Description

technical field [0001] The invention belongs to the technical field of medical image recognition, in particular to a method and device for detecting bone quality in CT images based on a convolutional neural network. Background technique [0002] Osteoporosis is a disease characterized by decreased bone mass and microarchitectural deterioration of bone tissue. This is a serious public health concern because of its potentially devastating consequences and high probability of fractures. From the patient's perspective, fractures and subsequent loss of mobility and autonomy often result in a reduced quality of life. Furthermore, osteoporotic vertebral fractures lasting 12 months carry an additional mortality rate of up to 20%, as they require hospitalization and subsequently increase the risk of other complications, such as pneumonia or thromboembolic disease from chronic immobilization. Moreover, because bone loss occurs insidiously and is initially asymptomatic, osteoporosis ...

AI Technical Summary

Problems solved by technology

In the prior art, the method of segmenting the lumbar spine from clinical CT images based on geometric and intensity features is used to detect wedge-shaped compression fractures, and the prototype algorithm for detecting osteoporotic vertebral fractures in conventional Doppler CT images of the chest and abdomen, and based on The machine learning method of the deep neural network framework still requires clinical additional auxiliary hardware and dedicated workflow for cases of bone loss without fracture

Images