Image recognition method and device

Abstract

The invention relates to an image recognition method. The method comprises steps: a three-dimensional imaging picture is acquired; the three-dimensional imaging picture is processed through a featureextraction model branch and picture features and a heat picture of the three-dimensional imaging picture are acquired, wherein the picture features are used for indicating features of a disease sourcearea in the three-dimensional imaging picture; and through a recognition model branch, the picture features and the heat picture are processed to acquire disease attributes corresponding to the three-dimensional imaging picture, wherein the disease attributes comprise benign or malignant attributes. As the constructed heat picture can indicate the global characteristics of the three-dimensional imaging picture, more pathology information can be carried, and thus, recognition based on the heat picture can improve the recognition accuracy.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of machine learning, and in particular to an image recognition method and device. Background technique [0002] With the continuous development of machine learning algorithms, the application of machine learning algorithms in the medical field is becoming more and more extensive. Automatic identification of benign and malignant diseases is a specific application of machine learning algorithms in the medical field. [0003] Taking lung cancer as an example, pulmonary nodules are a multi-system and multi-organ granulomatous disease of unknown etiology. In related technologies, benign and malignant lung cancers are usually identified according to benign and malignant pulmonary nodules. Specifically, in related technologies, two independent machine learning models are pre-trained, wherein the first machine learning model manually marks the CT (Computed Tomography, electronic computer tomogr...

AI Technical Summary

Problems solved by technology

[0006] The embodiment of the present invention provides an image recognition method and device, which can be used to solve the problems of low accuracy of the model trained in related technologies and low accuracy of judging benign and malignant diseases. The technical solution is as follows:

Images