pyradicomisc 提取组学特征

最新推荐文章于 2024-07-04 09:54:25 发布

原创最新推荐文章于 2024-07-04 09:54:25 发布 · 314 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#python #numpy #pandas

本文介绍了一个从400个病人的DICOM影像数据中提取放射组学特征的过程。使用了pyradiomics库进行特征提取，并对每个病人的影像数据进行了预处理，包括读取图像和掩膜数据、设置图像属性等步骤。最终将提取到的特征存储为CSV文件。

400个病人文件夹，每个病人文件夹下有2个文件：data.dcm ,mylabel.dcm


yaml_path = r'D:\caosh\作业\Params.yaml'  # 这是pyradicomisc用得配置文件
root=r'D:\caoshiwen\400dcm'
datas=glob.glob(r'D:\caosh\400dcm\*\data.dcm')
results = list()
indexs = list()

def prepare_images(data_path,mask_path):
    mask = sitk.ReadImage(mask_path)
    mask_arr = sitk.GetArrayFromImage(mask)
    # reader = sitk.ImageSeriesReader()
    # dicom_names = reader.GetGDCMSeriesFileNames(folderPath)
    # reader.SetFileNames(dicom_names)
    # image = reader.Execute()
    image = sitk.ReadImage(data_path)
    image_arr = sitk.GetArrayFromImage(image)  # Note: order:z, y, x !!
    # print(image_arr.shape)#(66, 512, 400)
    size = image.GetSize()#(400, 512, 66)
    origin = image.GetOrigin()  # order: x, y, z
    spacing = image.GetSpacing()  # order:x, y, z
    direction = image.GetDirection()

    pixelType = sitk.sitkInt8  # 注意这里是Int8
    image_new = sitk.Image(size, pixelType)
    mask_new = sitk.Image(size, pixelType)

    image_new = sitk.GetImageFromArray(image_arr)
    image_new.SetDirection(direction)
    image_new.SetSpacing(spacing)
    image_new.SetOrigin(origin)

    mask_new = sitk.GetImageFromArray(mask_arr)
    mask_new.SetDirection(direction)
    mask_new.SetSpacing(spacing)
    mask_new.SetOrigin(origin)
    return image_new,mask_new
def predict_features(image, mask, option_yaml_path):
    extractor = RadiomicsFeatureExtractor(option_yaml_path)
    return extractor.execute(image, mask)
m=1
for i in datas:
    m=m+1
    label=i.replace("data.dcm","segmentation_results.dcm")
    if os.path.exists(label) is False:
        label = i.replace("data.dcm", "mylabel.dcm")
    name=i.split("\\")[-2]
    print('now',m,name)
    indexs.append(str(name))
    images, masks=prepare_images(i,label)
    try:
         results.append(predict_features(images, masks, yaml_path))
    except:
        print('---->有问题',name)


df = pd.DataFrame(results)
df.drop(df.columns[list(range(22))], axis=1, inplace=True)  # drop the non feature
df.index = indexs
df.to_csv('400patients_radmioicc_features.csv')

结果：