Medical image data analysis method and medical image data analysis system

Abstract

The invention is applicable to the technical field of medical image processing and provides a medical image data analysis method. The medical image data analysis method includes the steps of acquiring medical image information; classifying the medical image information based on different parts; establishing medical models corresponding to the parts based on the classified medical image information; subjecting the medical models to training processing; storing model parameters. The invention further provides a medical image data analysis system correspondingly. The medical image data analysis method and the medical image data analysis system not only can achieve an image diagnosis function but also can provide later-period scheme recommendations.

Description

technical field [0001] The invention relates to the technical field of medical image processing, in particular to an analysis method and system for medical image data. Background technique [0002] The amount of medical image output data in medical institutions is very large, and image data often contains a large amount of potential information. At present, the medical market mainly relies on manual interpretation and analysis of medical images, which has low efficiency and limited information that can be mined, and cannot make full use of data resources. In recent years, with the rapid development of deep learning, various fields of machine learning, including computer vision, have been rapidly occupied by deep learning. Deep learning has achieved remarkable results in image recognition, speech recognition and other fields. As one of the important methods in machine learning, its powerful automatic feature extraction, complex model construction and image processing capabil...

AI Technical Summary

Problems solved by technology

[0003] Although the convolutional neural network has obvious advantages compared with the traditional feature extraction technology, it still has the following shortcomings: First, the current system does not have the function of recommending treatment plans, and only stays at the diagnostic level without considering Later more important countermeasures

Second, the convolutional neural network currently used is based on labeled data and implemented using the algorithm of error back propagation, which takes a long time to train and requires a large amount of labeled sample data

Moreover, as the number of hidden layers of the convolutional neural network increases, the feature learning ability of the first few layers of the model is particularly poor, and effective features cannot be learned.

[0004] In summary, there are obviously inconveniences and defects in the actual use of the existing technology, so it is necessary to improve

Images