Esophageal cancer detection method and system based on barium meal radiography image

Abstract

The invention provides an esophageal cancer detection method and system based on a barium meal contrast image. The method comprises the following steps: acquiring the barium meal contrast image of an esophagus to be detected; performing focus area detection on the barium meal contrast image according to the trained target detection model; wherein the target detection model adopts an improved Faster R-CNN, and a basic network of the target detection model is composed of a convolutional neural network carrying an attention mechanism; obtaining a feature map of the image according to the basic network, and enhancing the feature saliency of the feature map through an attention mechanism; and generating a region of interest and a feature vector of the region of interest according to the feature map, and obtaining detection information of the lesion region according to the feature vector of the region of interest. The attention mechanism is embedded in the basic network, the capability of obtaining the features of the region of interest of the target detection model is improved, and the esophageal cancer detection accuracy of the system is improved through fusion of multi-image and multi-body-position detection information.

Description

technical field [0001] The invention relates to the field of medical images, in particular to a method and system for detecting esophageal cancer using barium meal contrast images. Background technique [0002] Cancer is one of the most harmful diseases to human health today. Early diagnosis and treatment of many cancers can prevent and treat them. Early cancers often appear as small lesions on imaging, and it takes a lot of time for doctors to screen them. The use of computer-aided detection (Computer-Aided Detection, CAD) system can reduce the rate of missed detection of lesions by doctors and improve the efficiency of doctors' image reading. [0003] Esophageal cancer is a malignant tumor that occurs in the esophageal epithelial tissue. Early detection, early diagnosis and early treatment are still the most effective way to cure esophageal cancer. At present, the CAD system for esophageal cancer mainly focuses on the research of esophageal cancer cell images and esophag...

AI Technical Summary

Problems solved by technology

[0008] Although the CAD system using the Faster R-CNN network shows obvious advantages over other networks in the detection of barium meal contrast images, there is still room for further improvement in practical applications.

Images