Bone age evaluation method and device based on deep convolutional neural network and feature fusion, and computer readable storage medium

Abstract

The invention provides a bone age evaluation method and device based on a deep convolutional neural network and feature fusion and a computer readable storage medium, and the method comprises the following steps: collecting wrist bone X-ray films, and classifying the wrist bone X-ray films to form a wrist bone X-ray image set; randomly selecting a training image and a test image, marking an ROI region based on a TW scoring method, constructing a target detection neural network applying a YOLO model, and inputting the training image and the test image into the target detection neural network to form a deep convolutional neural network; training images are randomly selected and input to the deep convolutional neural network, global feature information of the training images is extracted by the deep convolutional neural network, local feature information of the training images is extracted based on the target detection neural network, and the global feature information and the local feature information are fused to obtain a bone age evaluation result. After the technical scheme is adopted, full automation of bone age evaluation can be realized, the network performance is improved, and the accuracy of bone age evaluation is further improved.

Description

technical field [0001] The present invention relates to the field of medical image processing, in particular to a bone age assessment method, device and computer-readable storage medium based on deep convolutional neural network and feature fusion. Background technique [0002] Skeletal age is referred to as bone age, which is an objective index to evaluate the bone development degree of adolescents and children and to measure biological age. Bone age assessment plays an important role in the growth and development of adolescents and children, disease diagnosis, judicial identification, and sports competitions. In terms of growth and development, bone age can give the biological age of children, estimate height, and evaluate whether the growth and development are normal. In terms of disease diagnosis, if the bone age is higher than the chronological age, it belongs to the state of early development, and the doctor will consider whether the individual suffers from diseases s...

AI Technical Summary

Problems solved by technology

[0005] Bone age reading is a labor-intensive task. Although the GP atlas method and the TW scoring method have their own characteristics, the GP atlas method is simple and the TW scoring method is accurate, but these methods have several common shortcomings. They all rely heavily on Due to the domain knowledge and experience of radiologists, the evaluation process takes a lot of time and manpower, and the time efficiency is low

Due to the strong subjectivity of manual interpretation, in many cases, even if the same expert evaluates the same X-ray image, the evaluation results will have certain differences at different times

At the same time, due to the higher acquisition cost of medical images and the need for professional radiologists for labeling, there are very limited datasets dedicated to bone age prediction with high-quality labels.

Most of the traditional automated evaluation methods require artificially designed features as input, which cannot meet the requirements of automation, and its performance is difficult to meet the requirements of practical applications.

In addition, although the bone age assessment model using deep learning methods can automatically extract features, less attention is paid to the ROI in bone age assessment, and the new deep learning theory can be further improved

Images