X-ray image processing method and system based on migration learning

Abstract

The application provides an X-ray image processing method and system based on transfer learning. The system includes: an X-ray image receiving module for receiving input X-ray image data; an X-ray image processing module that uses an optimization model to process all The X-ray image is processed and classified; the optimized model is obtained in the following manner: set up a VGG-16 convolutional neural network, and based on the number of X-ray image classification results, abandon the VGG-16 convolutional neural network. A fully connected layer, and optimize the training parameters to obtain a specific fully connected layer, and obtain a complete model; optimize the complete model to obtain an optimized model; an output module that receives the output data of the X-ray image processing module, and Output classification results. The solution of the present application has high accuracy of X-ray image processing and recognition for the abnormality of the large joint, and the system fitting is excellent, and the training efficiency is high.

Description

technical field [0001] The invention relates to the field of computer intelligent image processing, in particular to a method and system for processing bone-related X-ray effects. Background technique [0002] X-ray imaging examination is the main technical means and basis in the medical field, especially in the related subdivision fields such as bone observation. Taking the effect of X on hand bones as an example, in the presence of bone mutation or bone deformation caused by disease, the X-ray imaging changes are specific and can reflect the degree of bone mutation and the possible damage to various joints in the body. . In real skeleton X images, although skeletal deformation is relatively easy to observe and identify, its image manifestations are complex and diverse. There can even be dozens of manifestations of skeletal deformation caused by the same type, and most of them coexist. Therefore, the identification or classification of specific deformation types is extrem...

AI Technical Summary

Problems solved by technology

[0003] Based on the above existing problems, the existing X-ray image recognition and processing methods can no longer meet the current market demand. The identification and classification of the causes of bone deformation or variation with complex expressions has become an important task for X-ray image processing and recognition. Urgent problems to be solved in the field

Images