White matter high signal segmentation method, device and apparatus and storage medium

Abstract

The invention discloses a white matter high signal segmentation method, device and apparatus and a storage medium. The method comprises the steps of obtaining a to-be-segmented image; performing whitematter high signal segmentation on the to-be-segmented image based on a first segmentation network to obtain a white matter high signal label prediction result of the to-be-segmented image; and taking the to-be-segmented image and the white matter high signal label prediction result of the to-be-segmented image as input of a second segmentation network, and correcting the white matter high signallabel prediction result of the to-be-segmented image based on the second segmentation network to obtain a white matter high signal label correction result of the to-be-segmented image. By utilizing the technical scheme provided by the invention, the white matter high signal segmentation result with higher resolution, more accurate boundary and better spatial continuity can be obtained, and the problem of fuzzy boundary when the first segmentation network segments the white matter high signal is solved.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a white matter hyperintensity segmentation method, device, equipment and storage medium. Background technique [0002] Vascular white matter hyperintensities of presumed vascular origin (WMH), also known as leukoaraiosis, white matter lesion or white matter disease, refers to the fluid-attenuated inversion recovery sequence (Fluid-attenuated Inversion Recovery, FLAIR), there are multiple punctate, patchy or fusion hyperintensity phenomena around the bilateral lateral ventricles or in the subcortical white matter. White matter hyperintensities are common in older populations and in patients with small vessel disease or other neurological disorders, and are associated with increased risk of functional decline, dementia, and death. [0003] Clinically, white matter hyperintensities are generally assessed by visual rating scales. However, since MRI images often cont...

AI Technical Summary

Problems solved by technology

[0004] In recent years, deep learning technology has become a powerful alternative to supervised learning methods based on manually designed features. Deep learning technology has a powerful model capacity and can automatically learn to extract highly discriminative features. The performance achieved can be significantly better than manually designed features. However, in the prior art, the segmentation map generated by the white matter hyperintensity segmentation technology based on deep learning is often smooth, which cannot guarantee the continuity of the segmentation results in appearance and space, and the effect in the target boundary area is poor, so it is necessary to provide more effective solution

Images