Chest radiography lung field segmentation model establishment based on multi-scale feature fusion and segmentation method

Abstract

The invention discloses a chest radiography lung field segmentation model establishment based on multi-scale feature fusion and segmentation method. A segmentation model establishing method comprisesthe following steps: firstly, processing an X-ray chest radiograph; obtaining a preprocessed picture and a recoded mask picture, then constructing an X-ray chest radiography lung field segmentation network based on multi-scale convolution and a feature pyramid, and finally training the segmentation network by using the preprocessed picture as an input of the segmentation network and using the recoded mask picture as an output of the segmentation network to obtain a trained segmentation model; and based on the obtained segmentation model, preprocessing any to-be-processed X-ray chest radiograph, and inputting the preprocessed X-ray chest radiograph into the segmentation model to obtain a lung field segmentation result. According to the method, multi-resolution feature fusion is provided incombination with a feature pyramid theory, wherein the segmentation results with different resolutions can be fused, so that the segmentation effect is improved.

Description

technical field [0001] The invention belongs to the technical field of medical image analysis, and relates to a chest X-ray lung field segmentation model establishment and segmentation method based on multi-scale feature fusion. Background technique [0002] Chest X-ray (CXR) imaging is currently the most popular and available diagnostic tool for health monitoring and diagnosis of several lung diseases including pneumonia, tuberculosis, cancer, etc. But detecting these diseases from CXR is a very complex process and requires the involvement of radiologists. Moreover, millions of CXRs are generated every year. According to the National Health Service (UK), more than 22.9 million X-ray images were produced in the UK in 2017 / 18, accounting for all imaging equipment including magnetic resonance (MRI) and computed tomography (CT). 55.63%. The global shortage of radiologists suggests that this is a considerable diagnostic workload. [0003] An important step in the computer-ai...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the deficiencies in the prior art, the present invention provides a chest X-ray lung field segmentation model establishment and segmentation method based on multi-scale feature fusion, which solves the problem of under-segmentation (cavity) and edge of the lung field in the existing segmentation method. Segmentation is not accurate enough

Images