医学图像读取与绘制方法

医学图像读取与绘制程序，适用于.nii.gz(.nii)格式、.pkl格式、.npz格式

1.文件.nii格式读取显示：

'''
Author: <qilin>
Email: <EMAIL>
Date: 2024-02-20
Description:
This script converts nifti images into an image with
the corresponding.
'''


mode = 1



if mode == 1:

    '''
    读取nii文件内容
    '''
    import nibabel as nib
    import numpy as np
    
    
    file_path = r'F:/program/database/Multimodal/muregpro_pkl/us/test_nii/case000067.nii.gz'
    file_path = r'F:\program\database\Multimodal\muregpro\train\us_images\case000000.nii.gz'
    file_path = r'F:\program\database\Multimodal\OASIS_L2R_2021_task03\Test_nii\img0438.nii.gz'
    file_path = r'F:/program/database/Multimodal/muregpro/train/us_images/case000001.nii.gz'
    
    # 读取NIfTI文件
    nii_img = nib.load(file_path)
    
    # 获取头信息
    header = nii_img.header
    
    # 获取图像数据
    data = nii_img.get_fdata()
    
    # 如果数据是浮点数类型，将其转换为整型
    if data.dtype.kind == 'f':
        data = (data * 255).astype(np.uint8)
    
    # 输出标签和内容
    print("文件路径：", file_path)
    print("头信息：", header)
    print("图像数据：", data)
    print("图像维度： ",data.shape)


elif mode == 2:
    '''
    读取文件中元素频次
    '''
    
    import nibabel as nib
    import numpy as np
    
    # 读取NIfTI文件
    file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/train/mr_labels/case000000.nii.gz'
    img = nib.load(file_path)
    data = img.get_fdata()
    
    # 统计元素出现的频次
    unique_elements, counts = np.unique(data, return_counts=True)
    
    # 输出结果
    print("出现的元素及其频次：")
    for element, count in zip(unique_elements, counts):
        print(f"元素 {element}: 出现 {count} 次")


elif mode == 3:
    '''
    检查数据维度和最大最小值
    '''
    
    import nibabel as nib
    
    # 定义 NIfTI 文件路径
    nii_file_path = 'F:/文件/麒麟/program/python/TransMorph_Transformer_for_Medical_Image_Registration-main/Data/OASIS/Test_nii/img0438.nii.gz'
    label_file_path = 'F:/文件/麒麟/program/python/TransMorph_Transformer_for_Medical_Image_Registration-main/Data/muregpro/train/mr_labels/case000000.nii.gz'
    
    # 加载 NIfTI 文件
    nii_img = nib.load(label_file_path)
    
    # 获取数据数组
    data = nii_img.get_fdata()
    
    # 获取数据维度
    data_shape = data.shape
    
    max = 0
    min = 10000
    for i in range(len(data)):
        for j in range(len(data[i])):
            for k in range(len(data[i][j])):
                if data[i][j][k] > max:
                    max = data[i][j][k]
                elif data[i][j][k] < min:
                    min = data[i][j][k]
    print("max:",max)
    print("min:",min)
    
    print("数据维度：", data_shape)
    print("数据切片:",data[100][100])

elif mode == 4:
    '''
    显示平面切片
    '''
    
    import matplotlib.pyplot as plt
    import nibabel as nib
    
    # 读取 NIfTI 文件
    file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/val/mr_images/case000065.nii.gz'
    img = nib.load(file_path)
    data = img.get_fdata()
    
    # 可视化立体图像
    plt.figure(figsize=(8, 8))
    
    # 显示三个轴的切片
    plt.subplot(131)
    plt.imshow(data[:, :, data.shape[2] // 2], cmap='gray')
    plt.title('Axial')
    
    plt.subplot(132)
    plt.imshow(data[:, data.shape[1] // 2, :], cmap='gray')
    plt.title('Sagittal')
    
    plt.subplot(133)
    plt.imshow(data[data.shape[0] // 2, :, :], cmap='gray')
    plt.title('Coronal')
    
    plt.show()

elif mode == 5:
    '''
    逐层显示
    '''
    
    
    import numpy as np
    import nibabel as nib
    from mayavi import mlab
    
    # 定义 NIfTI 文件路径
    nii_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/val/mr_images/case000065.nii.gz'
    
    # 加载 NIfTI 文件
    nii_img = nib.load(nii_file_path)
    
    # 获取数据数组
    data = nii_img.get_fdata()
    
    # 将数据维度从 (120, 128, 128, 1) 转换为 (128, 128, 120)
    data = np.transpose(data.squeeze(), (1, 2, 0))
    
    # 创建 Mayavi 可视化窗口
    fig = mlab.figure(size=(800, 800))
    
    # 一次性加载整个3D图
    vol = mlab.volume_slice(data, colormap='gray')  # 灰度图
    # vol = mlab.volume_slice(data, opacity=0.5)     # 彩图
    
    # 显示结果
    mlab.show()

elif mode == 6:
    '''
    逐层显示(带标签)
    '''
    
    import numpy as np
    import nibabel as nib
    from mayavi import mlab
    
    # 定义 NIfTI 文件路径
    nii_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/train/us_images/case000000.nii.gz'
    label_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/train/us_labels/case000000.nii.gz'
    
    # 加载 NIfTI 文件
    nii_img = nib.load(nii_file_path)
    label_img = nib.load(label_file_path)
    
    # 获取数据数组
    data = nii_img.get_fdata()
    labels = label_img.get_fdata()
    
    # 将数据维度从 (120, 128, 128, 1) 转换为 (128, 128, 120)
    data = np.transpose(data.squeeze(), (1, 2, 0))
    # 将标签维度从 (120, 128, 128, 6) 转换为 (128, 128, 120, 6)
    labels = np.transpose(labels.squeeze(), (0, 1, 2, 3))
    
    # 创建 Mayavi 可视化窗口
    fig = mlab.figure(size=(800, 800))
    
    # 一次性加载整个3D图
    vol = mlab.volume_slice(data, colormap='gray')  # 灰度图
    # vol = mlab.volume_slice(data, opacity=0.5)     # 彩图
    
    # 找到标签值为255的位置
    x, y, z = np.where(labels[..., 0] == 255)  # 选择你想要查看的标签的索引，这里使用了第一个标签
    
    # 标注为红色
    mlab.points3d(x, y, z, color=(1, 0, 0), mode='cube', scale_factor=1, opacity=0.01)
    
    # 显示结果
    mlab.show()

elif mode == 7:
    '''
    立体图
    '''
    
    import numpy as np
    import nibabel as nib
    from mayavi import mlab
    
    # 定义 NIfTI 文件路径
    
    
    # nii_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/output/moved.nii.gz'
    # nii_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/val/us_images/case000069.nii.gz'
    
    # nii_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/val/us_images/case000066.nii.gz'
    # nii_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/val/us_images/case000067.nii.gz'
    # nii_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/dataTest/us_output/normal/moved.nii.gz'
    moving = 'F:/文件/麒麟/program/python/TransMorph_Transformer_for_Medical_Image_Registration-main/Data/OASIS/Test_nii/img0438.nii.gz'
    fixed = 'F:/文件/麒麟/program/python/TransMorph_Transformer_for_Medical_Image_Registration-main/Data/OASIS/Test_nii/img0439.nii.gz'
    moved = 'F:/文件/麒麟/program/python/TransMorph_Transformer_for_Medical_Image_Registration-main/Data/OASIS/Test_nii/img0440.nii.gz'
    
    # 加载 NIfTI 文件
    nii_img = nib.load(moving)
    nii_img_2 = nib.load(fixed)
    nii_img_3 = nib.load(moved)
    
    # 获取数据数组
    data = nii_img.get_fdata()
    data_2 = nii_img_2.get_fdata()
    data_3 = nii_img_3.get_fdata()
    
    # 将数据维度从 (120, 128, 128, 1) 转换为 (128, 128, 120)
    data = np.transpose(data.squeeze(), (1, 2, 0))
    data_2 = np.transpose(data_2.squeeze(), (1, 2, 0))
    data_3 = np.transpose(data_3.squeeze(), (1, 2, 0))
    
    # 创建 Mayavi 可视化窗口
    # fig = mlab.figure(size=(800, 800))
    # vol = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data), colormap='gray')  # 灰度图
    # vol = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data), opacity=0.9)     # 彩图 （opacity：0——1 不透明度）
    # 创建一个大小为800x800的图像窗口
    fig = mlab.figure(size=(800, 800))
    
    # 定义数据的形状
    shape = data.shape
    
    # 计算数据在x轴方向上的中心位置
    center_x = shape[0] // 6  # 左边1/3的中心位置
    
    # 在图像的左边1/3处显示数据
    vol = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data), colormap='gray')  # 灰度图
    vol.actor.actor.position = np.array([-center_x, 0, 0])
    
    
    
    fig_2 = mlab.figure(size=(800, 800))
    # vol_2 = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data_2), colormap='gray')  # 灰度图
    vol_2 = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data_2), opacity=0.9)     # 彩图 （opacity：0——1 不透明度）
    
    fig_3 = mlab.figure(size=(800, 800))
    # vol_3 = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data_3), colormap='gray')  # 灰度图
    vol_3 = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data_3), opacity=0.9)     # 彩图 （opacity：0——1 不透明度）
    
    # 显示结果
    mlab.show()

elif mode == 8:
    '''
    立体图(带标签)
    '''
    
    import numpy as np
    import nibabel as nib
    from mayavi import mlab
    
    # 定义 NIfTI 文件路径
    nii_file_path = 'F:/文件/麒麟/program/python/TransMorph_Transformer_for_Medical_Image_Registration-main/Data/muregpro/train/mr_images/case000001.nii.gz'
    label_file_path = 'F:/文件/麒麟/program/python/TransMorph_Transformer_for_Medical_Image_Registration-main/Data/muregpro/train/mr_labels/case000001.nii.gz'
    
    # nii_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/val/us_images/case000069.nii.gz'
    # label_file_path = '/home/adminer/qilin/program/voxelmorph-dev/data/muregpro/val/us_labels/case000069.nii.gz'
    
    # 加载 NIfTI 文件
    nii_img = nib.load(nii_file_path)
    label_img = nib.load(label_file_path)
    
    # 获取数据数组
    data = nii_img.get_fdata()
    labels = label_img.get_fdata()
    
    # 将数据维度从 (120, 128, 128, 1) 转换为 (128, 128, 120)
    data = np.transpose(data.squeeze(), (1, 2, 0))
    # 将标签维度从 (120, 128, 128, 6) 转换为 (6, 128, 128, 120)
    labels = np.transpose(labels.squeeze(), (0, 1, 2, 3))
    
    # 创建 Mayavi 可视化窗口
    fig = mlab.figure(size=(800, 800))
    
    # 创建等值面
    vol = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data), colormap='gray', opacity=0.9)  # 灰度图
    # vol = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data), opacity=0.9)  # 彩图 （opacity：0——1 不透明度）
    
    # 找到标签值为255的位置,标注为红色
    x, y, z = np.where(labels[..., 0] == 255)  # 选择你想要查看的标签的索引，这里使用了第一个标签
    mlab.points3d(x, y, z, color=(1, 0, 0), mode='cube', scale_factor=1, opacity=0.1)
    
    x, y, z = np.where(labels[..., 1] == 255)  # 选择你想要查看的标签的索引，这里使用了第一个标签
    mlab.points3d(x, y, z, color=(0, 1, 0), mode='cube', scale_factor=1, opacity=0.5)
    
    x, y, z = np.where(labels[..., 2] == 255)  # 选择你想要查看的标签的索引，这里使用了第一个标签
    mlab.points3d(x, y, z, color=(0, 0, 1), mode='cube', scale_factor=1, opacity=0.5)
    
    x, y, z = np.where(labels[..., 3] == 255)  # 选择你想要查看的标签的索引，这里使用了第一个标签
    mlab.points3d(x, y, z, color=(1, 1, 0), mode='cube', scale_factor=1, opacity=0.5)
    
    x, y, z = np.where(labels[..., 4] == 255)  # 选择你想要查看的标签的索引，这里使用了第一个标签
    mlab.points3d(x, y, z, color=(1, 0, 1), mode='cube', scale_factor=1, opacity=0.5)
    
    x, y, z = np.where(labels[..., 5] == 255)  # 选择你想要查看的标签的索引，这里使用了第一个标签
    mlab.points3d(x, y, z, color=(0, 1, 1), mode='cube', scale_factor=1, opacity=0.5)
    # 显示结果
    mlab.show()

2.文件.pkl格式读取与显示

'''
Author: <qilin>
Email: <EMAIL>
Date: 2024-02-28
'''
import pickle
import numpy as np
import nibabel as nib
from mayavi import mlab
import matplotlib.pyplot as plt

'''
读取pkl文件内容
'''

def read_pkl(path):
    # 加载.pkl文件
    with open(path, 'rb') as file:
        data = pickle.load(file)

    # 将数据转换为Numpy数组
    data = np.array(data)

    # 获取数据的维度
    data_shape = data.shape

    print("数据维度：", data_shape)


'''
数据滤波
'''
def filter_y_and_z(y, z):
    # 手动统计y数组中每个元素的出现次数
    y_count = {}
    for num in y:
        if num in y_count:
            y_count[num] += 1
        else:
            y_count[num] = 1

    # 找出出现次数大于等于3次的元素
    y_to_keep = [num for num, count in y_count.items() if count >= 3]

    # 将y数组中出现次数大于等于3次的元素所对应的z数组中的数据存储到字典中
    y_z_mapping = {}
    for idx, num in enumerate(y):
        if num in y_to_keep:
            if num not in y_z_mapping:
                y_z_mapping[num] = []
            y_z_mapping[num].append(z[idx])

    # 对字典中的每个元素，只保留最大值和最小值
    for num, z_values in y_z_mapping.items():
        max_z = max(z_values)
        min_z = min(z_values)
        y_z_mapping[num] = [max_z, min_z]

    # 将不在y_to_keep中的y数组元素对应的z数组中数据置为None
    filtered_y = []
    filtered_z = []
    for idx, num in enumerate(y):
        if num in y_to_keep:
            filtered_y.append(num)
            filtered_z.append(y_z_mapping[num][0] if z[idx] == max(y_z_mapping[num]) else y_z_mapping[num][1])

    return filtered_y, filtered_z
'''
标签突出
'''
def display_seg_3D(data):
    x, y, z = np.where(data == 1)
    mlab.points3d(x, y, z, color=(1, 0, 0), mode='cube', scale_factor=1, opacity=0.1) # 红

    x, y, z = np.where(data == 2)
    mlab.points3d(x, y, z, color=(0, 1, 0), mode='cube', scale_factor=1, opacity=0.5) # 绿

    x, y, z = np.where(data == 3)
    mlab.points3d(x, y, z, color=(0, 0, 1), mode='cube', scale_factor=1, opacity=0.5) # 蓝

    x, y, z = np.where(data == 4)
    mlab.points3d(x, y, z, color=(1, 1, 0), mode='cube', scale_factor=1, opacity=0.5) # 黄

    x, y, z = np.where(data == 5)
    mlab.points3d(x, y, z, color=(1, 0, 1), mode='cube', scale_factor=1, opacity=0.5) # 粉

    x, y, z = np.where(data == 6)
    mlab.points3d(x, y, z, color=(0, 1, 1), mode='cube', scale_factor=1, opacity=0.5) # 青

def display_seg_Section_1(y, z):
    return y,z,y,z
def display_seg_Section_2(y, z):
    # 重新排序
    sorted_indices = sorted(range(len(y)), key=lambda k: y[k])
    z = [z[i] for i in sorted_indices]
    y = sorted(y)

    num_x1 = []
    num_x2 = []
    num_y1 = []  # 大rang
    num_y2 = []  # 小
    j = -1
    for i in range(len(y)):
        if y[i] in num_x2:
            if z[i] > num_y1[j]:
                num_y1[j] = z[i]
            elif z[i] < num_y2[j]:
                num_y2[j] = z[i]
        else:
            j += 1
            num_x1.append(y[i])
            num_y1.append(z[i])
            num_x2.append(y[i])
            num_y2.append(z[i])
    # print(x,y,z)
    # print(len(x),len(y),len(z))
    # print(num_x1,num_y1)
    # print(num_x2,num_y2)
    # print(len(x),len(num_x1),len(num_y1),len(num_x2),len(num_y2))
    return num_x1, num_y1,num_x2, num_y2,

def display_seg_sectionAll(data,x_section):
    global mode
    global thickness
    if mode == 1:
        Display = display_seg_Section_1
    elif mode == 2:
        Display = display_seg_Section_2

    y, z = np.where(data[x_section] == 1)
    a,b,c,d = Display(y, z)
    plt.scatter(a, b, color=(1, 0, 0), s=thickness) # 红
    plt.scatter(c, d, color=(1, 0, 0), s=thickness)

    y, z = np.where(data[x_section] == 2)
    a, b, c, d = Display(y, z)
    plt.scatter(a, b, color=(0, 1, 0), s=thickness) # 绿
    plt.scatter(c, d, color=(0, 1, 0), s=thickness)

    # y, z = np.where(data[x_section] == 3)
    # a, b, c, d = Display(y, z)
    # plt.scatter(a, b, color=(0, 0, 1), s=thickness) # 蓝
    # plt.scatter(c, d, color=(0, 0, 1), s=thickness)
    #
    # y, z = np.where(data[x_section] == 4)
    # a, b, c, d = Display(y, z)
    # plt.scatter(a, b, color=(1, 1, 0), s=thickness) # 黄
    # plt.scatter(c, d, color=(1, 1, 0), s=thickness)
    #
    # y, z = np.where(data[x_section] == 5)
    # a, b, c, d = Display(y, z)
    # plt.scatter(a, b, color=(1, 0, 1), s=thickness) # 粉
    # plt.scatter(c, d, color=(1, 0, 1), s=thickness)
    #
    # y, z = np.where(data[x_section] == 6)
    # a, b, c, d = Display(y, z)
    # plt.scatter(a, b, color=(0, 1, 1), s=thickness) # 青
    # plt.scatter(c, d, color=(0, 1, 1), s=thickness)


'''
立体图(带标签)
'''
def display_3D(data_1,data_3,data_2,data_4):
    # 显示mr图像
    fig = mlab.figure(size=(800, 800))
    vol = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data_1), colormap='gray', opacity=0.9)  # 灰度图
    # vol = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data_1), opacity=0.9)  # 彩图 （opacity：0——1 不透明度）
    display_seg_3D(data_2)
    mlab.show()

    # 显示us图像
    fig_2 = mlab.figure(size=(800, 800))
    vol_2 = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data_3), colormap='gray', opacity=0.9)  # 灰度图
    # vol_2 = mlab.pipeline.iso_surface(mlab.pipeline.scalar_field(data_1), opacity=0.9)  # 彩图 （opacity：0——1 不透明度）
    display_seg_3D(data_4)
    mlab.show()

'''
剖面图(带标签)
'''
def diaplay_Section(data_1,data_3,data_2,data_4):
    # 可视化立体图像
    plt.figure(figsize=(8, 8))

    x_section = [data_1.shape[0] // 4 * 1,data_1.shape[0] // 4 * 2,data_1.shape[0] // 4 * 3]
    x_section_2 = [data_3.shape[0] // 4 * 1,data_3.shape[0] // 4 * 2,data_3.shape[0] // 4 * 3]

    # 显示切片
    plt.subplot(231)
    plt.imshow(data_1[x_section[0], :, :], cmap='gray')
    display_seg_sectionAll(data_2,x_section[0])
    plt.title('Coronal')

    plt.subplot(232)
    plt.imshow(data_1[x_section[1], :, :], cmap='gray')
    display_seg_sectionAll(data_2,x_section[1])
    plt.title('Coronal')

    plt.subplot(233)
    plt.imshow(data_1[x_section[2], :, :], cmap='gray')
    display_seg_sectionAll(data_2,x_section[2])
    plt.title('Coronal')

    plt.subplot(234)
    plt.imshow(data_3[x_section_2[0], :, :], cmap='gray')
    display_seg_sectionAll(data_4,x_section_2[0])
    plt.title('Coronal')

    plt.subplot(235)
    plt.imshow(data_3[x_section_2[1], :, :], cmap='gray')
    display_seg_sectionAll(data_4,x_section_2[1])
    plt.title('Coronal')

    plt.subplot(236)
    plt.imshow(data_3[x_section_2[2], :, :], cmap='gray')
    display_seg_sectionAll(data_4,x_section_2[2])
    plt.title('Coronal')

    plt.show()

'''
主函数
'''
# pkl_file_path = 'F:\program\database\Multimodal\OASIS_L2R_2021_task03/Test/p_0438_0439.pkl'
pkl_file_path = '/home/adminer/qilin/data/muregpro_ql/mr/train/case000001.pkl'
pkl_file_path_2 = '/home/adminer/qilin/data/muregpro_ql/mr/val/case000066.pkl'

mode = 1  # 显示模式 1:全显示 2:边界显示
thickness = 0.5
with open(pkl_file_path, 'rb') as f:
    data_1,data_2 = pickle.load(f)  # data_1,data_2分别是mr图像的img与seg
with open(pkl_file_path_2, 'rb') as f:
    data_3,data_4 = pickle.load(f)  # data_1,data_2分别是mr图像的img与seg
data_1 = np.transpose(data_1.squeeze(), (1, 2, 0))
data_3 = np.transpose(data_3.squeeze(), (1, 2, 0))

# data_2 = np.transpose(data_2.squeeze(), (1, 2, 0))
# data_4 = np.transpose(data_4.squeeze(), (1, 2, 0))


display_3D(data_1,data_3,data_2,data_4)
# diaplay_Section(data_1,data_3,data_2,data_4)mureg

3.文件.npz格式读取与绘制

'''
Author: <qilin>
Email: <EMAIL>
Date: 2024-01-19
Description:
This script converts npz files to an image file that can be used for visualization.
'''

import numpy as np
import matplotlib.pyplot as plt

mode = 4  # 可选模式: 1.折线图绘制 2.输出npz文件内数组名 3.平面切片绘图 4.立体图像绘制
file_name = 'F:/文件/麒麟/program/python/hypermorph-dev/data/dataTest/train1.npz'
array_name = 'vol'
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/dataTest/train1.npz' ['vol', 'seg', 'train_avg']
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/dataTest/train2.npz' ['vol', 'seg', 'train_avg']
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/atlas.npz'           ['vol', 'seg', 'train_avg']
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/generated_uncond_atlas.npz' ['vol']
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/labels.npz'   ['labels']
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/prob_atlas.npz' ['vol']
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/prob_atlas_mapping.npz' ['mapping']
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/prob_atlas_T1_stats.npz' ['init_mu', 'init_sigma']
# 'F:/文件/麒麟/program/python/hypermorph-dev/data/test_scan.npz' ['vol', 'seg']
'''

3. 'F:/文件/麒麟/program/python/hypermorph-dev/data/atlas.npz'
键值对：['vol', 'seg', 'train_avg']
含义：这个 .npz 文件包含一个参考图谱（atlas）。'vol' 键对应的值是图谱的体积数据，'seg' 键对应的值是图谱的分割数据，'train_avg' 键对应的值是训练平均数据。

4.'F:/文件/麒麟/program/python/hypermorph-dev/data/generated_uncond_atlas.npz'
键值对：['vol']
含义：这个 .npz 文件包含一个生成的无条件图谱（generated unconditional atlas）。'vol' 键对应的值是图谱的体积数据。

5.'F:/文件/麒麟/program/python/hypermorph-dev/data/labels.npz'
键值对：['labels']
含义：这个 .npz 文件包含标签数据（labels）。'labels' 键对应的值是标签数据。

6.'F:/文件/麒麟/program/python/hypermorph-dev/data/prob_atlas.npz'
键值对：['vol']
含义：这个 .npz 文件包含概率图谱（probabilistic atlas）。'vol' 键对应的值是概率图谱的体积数据。

7.'F:/文件/麒麟/program/python/hypermorph-dev/data/prob_atlas_mapping.npz'
键值对：['mapping']
含义：这个 .npz 文件包含概率图谱的映射数据（mapping）。'mapping' 键对应的值是概率图谱的映射数据。

8.'F:/文件/麒麟/program/python/hypermorph-dev/data/prob_atlas_T1_stats.npz'
键值对：['init_mu', 'init_sigma']
含义：这个 .npz 文件包含概率图谱 T1 统计数据。'init_mu' 键对应的值是概率图谱的均值数据，'init_sigma' 键对应的值是概率图谱的标准差数据。

9.'F:/文件/麒麟/program/python/hypermorph-dev/data/test_scan.npz'
键值对：['vol', 'seg']
含义：这个 .npz 文件包含测试扫描数据。'vol' 键对应的值是扫描的体积数据，'seg' 键对应的值是扫描的分割数据。
'''


'''
折线图绘制
'''
if mode == 1:
    # 加载.npz文件 ['vol', 'seg', 'train_avg']
    # data = np.load('F:/文件/麒麟/实验室/医学图像/论文/开源库/hypermorph-dev/data/dataTest/train1.npz')
    data = np.load(file_name)

    # 查看文件中保存的数组名称
    array_names = data.files
    print("Arrays in the .npz file:", array_names)

    # 遍历数组并进行可视化
    for name in array_names:
        array = data[name]
        print("Array:", name)
        print("Shape:", array.shape)
        print("Data:", array)

        # 可视化数组
        plt.figure()
        if array.ndim == 1:
            plt.plot(array)
        elif array.ndim == 2:
            plt.imshow(array, cmap='gray')
        else:
            print("Skipping visualization for array", name, "due to high dimensions.")
            continue

        plt.title(name)
        plt.show()

    '''
    输出npz文件内数组名
    '''
elif mode == 2:
    # 加载.npz文件
    data = np.load(file_name)

    # 查看文件中保存的数组名称
    array_names = data.files
    print("Arrays in the .npz file:", array_names)

    '''
    平面切片绘图
    '''
elif mode == 3:

    # 加载.npz文件 ['vol', 'seg', 'train_avg']
    data = np.load(file_name)

    # 获取图像数组
    image = data[array_name]

    # 选择切片索引
    slice_index = 80

    # 提取特定切片
    slice_image = image[slice_index, :, :]

    # 可视化切片图像
    plt.imshow(slice_image, cmap='gray')
    plt.axis('off')  # 关闭坐标轴
    plt.show()

    '''
    立体图像绘制
    '''
elif mode == 4:
    import numpy as np
    from mayavi import mlab

    # 加载.npz文件
    data = np.load(file_name)

    # 获取图像数组
    image = data[array_name]

    # 创建Mayavi场景
    mlab.figure(bgcolor=(1, 1, 1))

    # 创建体数据对象
    vol = mlab.pipeline.volume(mlab.pipeline.scalar_field(image))

    # 显示体数据
    mlab.show()