Automatic organ sketching method and device based on deep learning and radiotherapy equipment

Abstract

The invention discloses an automatic organ sketching method and device based on deep learning and radiotherapy equipment. The automatic organ sketching method specifically comprises the following steps: S1, acquiring a CT image of a patient; s2, preprocessing the CT image acquired in the step S1; s3, inputting the preprocessed CT image obtained in S2 into a large organ deep learning network model to obtain a large organ sketching result; s4, integrating the preprocessed CT image obtained in S2 and the delineation result of the large organ obtained in S3, and inputting the integrated image and the integrated image into a small organ deep learning network model to obtain a delineation result of the small organ; and S5, generating an organ sketching of the patient by combining the sketching result of the large organ obtained in the S3 and the sketching result of the small organ obtained in the S4, and realizing automatic sketching of the organ. According to the method, the neural network is used to predict the simple large organ sketching, and then the CT image and the sketching result of the large organ are integrated to predict the difficult small organ sketching, so that the sketching precision of the small organ is greatly improved.

Description

technical field [0001] The invention belongs to the technical field of radiotherapy, and in particular relates to a deep learning-based automatic organ delineation method, device and radiotherapy equipment. Background technique [0002] In radiation therapy, the delineation of organs at risk is a very important step of preparation, but the current traditional method is to let radiologists manually delineate the organs at risk of patients. The work efficiency of radiotherapy physicians is low; moreover, due to the repetitiveness of manual delineation, doctors are more likely to make delineation errors, thus introducing unnecessary human delineation deviations. [0003] In order to solve the above problems, researchers have turned their attention to the increasingly popular deep learning. Convolutional Neural Network (CNN) is the most common and effective method in the field of deep learning and image processing. Through a neural network algorithm, we can easily use thousand...

AI Technical Summary

Problems solved by technology

[0004] However, the existing automatic delineation algorithms based on convolutional neural networks are faced with the following problems: low accuracy and poor effect when predicting organs, especially when the predicted organs are small, the problem is more significant

Images