A medical image feature recognition method and device based on space-frequency domain feature fusion

Abstract

The present invention relates to the technical field of medical image-aided diagnosis, and in particular to a method and device for identifying medical image features based on feature fusion in the space-frequency domain. The method includes: acquiring medical images to be diagnosed; processing to obtain an airspace standard medical image; performing image preprocessing on the medical image to be diagnosed in the frequency domain to obtain a frequency domain standard medical image; fusing the airspace standard medical image and the frequency domain standard medical image to obtain Standard medical image features: input the standard medical image features into the trained convolutional neural network model to obtain the predicted value of the corresponding waiting medical image, if the predicted value is greater than the set diagnostic threshold, the diagnosis result is determined to be positive , otherwise it is judged that the diagnosis result is negative. The invention improves the efficiency of disease diagnosis for patients based on medical images, and improves the consistency and accuracy of medical image diagnosis.

Description

technical field [0001] The present invention relates to the technical field of medical image aided diagnosis, in particular to a method and device for identifying medical image features based on space-frequency domain feature fusion. Background technique [0002] In the medical imaging diagnosis industry, there are problems such as a large gap in high-end talents caused by the long training cycle of professional doctors, and low work efficiency caused by the time-consuming reading of films by professional doctors. In order to solve these problems, computer-aided technology is applied in medical imaging diagnosis. In the early days, the method of manually defining image features and then classifying them based on machine learning was often used. With the rapid development of deep learning, the current mainstream method is to use the neural network model to automatically extract and select image features and classify them. The steps are: first preprocess the medical images u...

AI Technical Summary

Problems solved by technology

[0003] However, in the existing methods of using deep learning for medical image diagnosis, in order to avoid the problem of a large number of neural network model parameters and low training efficiency caused by excessive medical image resolution, the image will be artificially greatly reduced in the preprocessing stage. resolution

This makes the extracted visual features lose the fine-grained information on the original medical images, which affects the diagnostic effect of medical images.

Images