本筆記來(lái)源于B站Up主: 有Li 的影像組學(xué)系列教學(xué)視頻
本節(jié)(32)主要介紹: 使用SimpleITK進(jìn)行N4偏置場(chǎng)校正

N4偏置場(chǎng)的定義和產(chǎn)生原因

MR scans often display intensity non-uniformities due to variations in the magnetic field. So, one part of an image might appear lighter or darker when visualized, solely because of variations in the magnetic field. The map of these variations is called the bias field. The bias field can cause problems for a classifier as the variations in signal intensity are not due to any anatomical differences.

代碼實(shí)現(xiàn)

import SimpleITK as sitk
imagePath = 'C:/Users/RONG/Desktop/Images/001/Lung_image.nrrd'

input_image = sitk.ReadImage(imagePath)
mask_image = sitk.OtsuThreshold(input_image,0,1,200)
input_image = sitk.Cast(input_image, sitk.sitkFloat32)
corrector = sitk.N4BiasFieldCorrectionImageFilter()
output_image = corrector.Execute(input_image,mask_image)
output_image = sitk.Cast(output_image, sitk.sitkInt16)
sitk.WriteImage(output_image, 'C:/Users/RONG/Desktop/Images/001/Lung_N4.nii.gz')
sitk.WriteImage(mask_image, 'C:/Users/RONG/Desktop/Images/001/Lung_mask.nii.gz')

知識(shí)擴(kuò)展

Otsu Thresholding
外文原帖講解較為詳細(xì)
中文版點(diǎn)這里

greyscale.JPG

像素類型.png