Artificial intelligence medical image automatic diagnosis system and method

Abstract

The embodiment of the invention provides an artificial intelligence medical image automatic diagnosis system and method. The method comprises the steps: collecting medical microscope images and corresponding diagnosis data; labeling the medical microscope images to obtain labeling data corresponding to the medical microscope images; and constructing a training set and a test set based on the diagnosis data and the annotation data corresponding to the medical microscope images, and performing training based on a deep learning model to obtain an optimal AI classification model and an optimal AIsemantic segmentation model, thereby realizing automatic diagnosis of the medical microscope images of a detection sample. According to the embodiment of the invention, human resources can be effectively saved, diagnosis time is shortened, and diagnosis accuracy is improved.

Description

technical field [0001] The present disclosure relates to the technical field of medical image processing, in particular, to an artificial intelligence medical image automatic diagnosis system and method. Background technique [0002] Female reproductive tract infection is a common disease in female gynecology, and it is also a global social and public health problem. Bacterial vaginosis (BV) is the most common disease of female reproductive tract infection. In normal vaginal flora, Lactobacillus predominates. Lactobacilli are Gram-positive large bacteria that are microaerophilic, but grow better in an anaerobic environment. A coordinated and balanced state is maintained between the host and the flora and between the flora and the flora. Bacterial vaginosis will easily form if the Lactobacillus in the female vagina is greatly reduced and Gardnerella, Bacteroides or Campylobacter grows in large quantities. Clinically, according to the distribution density of Lactobacillus,...

AI Technical Summary

Problems solved by technology

However, in medical image diagnosis, the number of diseases processed by medical images is very large, and the degree of difference between each disease is also very large. For most diseases, it is still at the level of manual diagnosis.

In particular, the diagnosis of gynecological microecological microscope images of the above-mentioned common gynecological diseases can only be performed manually at present, and there is no relevant intelligent diagnosis technology that can assist in the diagnosis. Samples, through the observation and analysis of the microscope images of the samples, and using the medical experience of the inspectors to make a diagnosis conclusion, this process consumes a lot of human resources, the diagnosis efficiency is low, and the diagnosis accuracy is relatively low

Images