Medical image recognition method and system based on deep learning

Abstract

The invention relates to the technical field of image recognition, and discloses a medical image recognition method based on deep learning, and the method comprises the steps: obtaining a to-be-recognized medical image, and carrying out binarization of the image through a binarization method based on a threshold value; performing noise reduction processing on the binarized image by using a medianfiltering method; performing image enhancement on the denoised image by using a gamma correction algorithm; positioning effective information of the enhanced medical image by using an image parameterdetection algorithm, wherein the effective information in the medical image comprises cell image information and organ image information; solving a multi-threshold segmentation target function based on a heuristic algorithm, and performing image segmentation on the detected and positioned medical image according to a solving result; and inputting the segmented medical image into a pre-trained FT-Densent model, and carrying out identification of the medical image. The invention further provides a medical image recognition system based on deep learning. According to the invention, medical imagerecognition is realized.

Description

technical field [0001] The present invention relates to the technical field of image recognition, in particular to a medical image recognition method and system based on deep learning. Background technique [0002] With the rapid development of the Internet and the continuous maturity and progress of computer technology, image processing technology in computer technology has been widely used. With further research in machine learning, its application in medical images in the field of biomedical engineering is becoming more and more important. Nowadays, the incidence of various diseases continues to rise. How to identify medical images of patients and provide feasible diagnosis opinions based on the identification results has become a hot topic in current research. [0003] Existing medical image recognition technology mainly obtains the cell region of the medical image by performing threshold segmentation on the medical image for recognition. Threshold segmentation can be ...

AI Technical Summary

Problems solved by technology

Threshold segmentation can be considered to include the two dimensions of objective function calculation and optimal value solution method. The key to threshold segmentation can be considered as the determination of the objective function and the calculation method and optimal solution solution method. When the number of thresholds increases and When the dimensionality of the threshold representation increases, that is, when the dimensionality of the decision space for objective function optimization increases, the computational complexity will increase dramatically, as with many optimization problems. Existing solutions based on exhaustive methods and classical mathematical model methods require The calculation time often cannot meet the real-time and practical requirements of image threshold segmentation

[0004] As the depth of the network increases, the propagation of gradients in the training of existing medical image recognition models becomes more difficult

Since the gradient propagation is calculated by multiplication, when the model becomes too deep, the number of multiplications becomes more and more, and the problem of gradient disappearance and gradient explosion will appear. If the product is too small, the gradient will disappear, and the product When it is too large, it will cause the gradient to explode

And as the depth of the model increases, the accuracy of the model becomes saturated, no longer increases, and then declines rapidly, resulting in model performance degradation

Images