def load_patch(dir_path, return_roi=False, parent_dir=None):
phasenames = ['NC', 'ART', 'PV']
mhd_images = []
mask_images = []
for phasename in phasenames:
# print os.path.join(parent_dir, basename, phasename + '_Image*.mhd')
# print os.path.join(dir_path, phasename + '_Image*.mhd')
image_path = glob(os.path.join(dir_path, phasename + '_Image*.mhd'))[0]
mask_path = os.path.join(dir_path, phasename + '_Registration.mhd')
mhd_image = read_mhd_image(image_path,
rejust=False) # 因为存储的是ｎｐｙ格式，所以不进行窗宽窗位的调整
mhd_image = np.squeeze(mhd_image)
# show_image(mhd_image)
mask_image = read_mhd_image(mask_path)
mask_image = np.squeeze(mask_image)
[xmin, xmax, ymin, ymax] = get_boundingbox(mask_image)
# xmin -= 15
# xmax += 15
# ymin -= 15
# ymax += 15
mask_image = mask_image[xmin:xmax, ymin:ymax]
mhd_image = mhd_image[xmin:xmax, ymin:ymax]
mhd_image[mask_image != 1] = 0
mask_images.append(mask_image)
mhd_images.append(mhd_image)
mhd_images = convert2depthlaster(mhd_images)
return mhd_images
def extract_patch(dir_name, suffix_name, patch_size, patch_step=1, flatten=False):
'''
提取指定类型病灶的ｐａｔｃｈ
:param patch_size: 提取ｐａｔｃｈ的大小
:param dir_name: 目前所有病例的存储路径
:param suffix_name: 指定的病灶类型的后缀，比如说cyst 就是０
:param patch_step: 提取ｐａｔｃｈ的步长
:return: patch_arr (图像的个数, patch的个数)
'''
count = 0
names = os.listdir(dir_name)
patches_arr = []
paths = []
for name in names:
if name.endswith(suffix_name):
# 只提取指定类型病灶的ｐａｔｃｈ
mask_images = []
mhd_images = []
paths.append(os.path.join(dir_name, name))
for phasename in phasenames:
image_path = glob(os.path.join(dir_name, name, phasename + '_Image*.mhd'))[0]
mask_path = os.path.join(dir_name, name, phasename + '_Registration.mhd')
mhd_image = read_mhd_image(image_path, rejust=True)
mhd_image = np.squeeze(mhd_image)
# show_image(mhd_image)
mask_image = read_mhd_image(mask_path)
mask_image = np.squeeze(mask_image)
[xmin, xmax, ymin, ymax] = get_boundingbox(mask_image)
mask_image = mask_image[xmin: xmax, ymin: ymax]
mhd_image = mhd_image[xmin: xmax, ymin: ymax]
mhd_image[mask_image != 1] = 0
mask_images.append(mask_image)
mhd_images.append(mhd_image)
# show_image(mhd_image)
mask_images = convert2depthlaster(mask_images)
mhd_images = convert2depthlaster(mhd_images)
# show_image(mhd_images)
count += 1
[width, height, depth] = list(np.shape(mhd_images))
patch_count = 1
# if width * height >= 400:
# patch_step = int(math.sqrt(width * height / 100))
patches = []
for i in range(patch_size / 2, width - patch_size / 2, patch_step):
for j in range(patch_size / 2, height - patch_size / 2, patch_step):
cur_patch = mhd_images[i - patch_size / 2:i + patch_size / 2,
j - patch_size / 2: j + patch_size / 2, :]
if (np.sum(
mask_images[i - patch_size / 2:i + patch_size / 2, j - patch_size / 2: j + patch_size / 2,
:]) / ((patch_size - 1) * (patch_size - 1) * 3)) < 0.9:
continue
patch_count += 1
if flatten:
patches.append(np.array(cur_patch).flatten())
else:
patches.append(cur_patch)
if patch_count == 1:
continue
patches_arr.append(patches)
return np.array(patches_arr)
def extract_patch_npy(dir_name, suffix_name, save_dir, patch_size, patch_step=1):
'''
提取指定类型病灶的ｐａｔｃｈ 保存原始像素值，存成ｎｐｙ的格式
:param patch_size: 提取ｐａｔｃｈ的大小
:param dir_name: 目前所有病例的存储路径
:param suffix_name: 指定的病灶类型的后缀，比如说cyst 就是０
:param save_dir: 提取得到的ｐａｔｃｈ的存储路径
:param patch_step: 提取ｐａｔｃｈ的步长
:return: None
'''
count = 0
names = os.listdir(dir_name)
for name in names:
if name.endswith(suffix_name):
# 只提取指定类型病灶的ｐａｔｃｈ
mask_images = []
mhd_images = []
for phasename in phasenames:
image_path = glob(os.path.join(dir_name, name, phasename + '_Image*.mhd'))[0]
mask_path = os.path.join(dir_name, name, phasename + '_Registration.mhd')
mhd_image = read_mhd_image(image_path)
mhd_image = np.squeeze(mhd_image)
# show_image(mhd_image)
mask_image = read_mhd_image(mask_path)
mask_image = np.squeeze(mask_image)
[xmin, xmax, ymin, ymax] = get_boundingbox(mask_image)
mask_image = mask_image[xmin: xmax, ymin: ymax]
mhd_image = mhd_image[xmin: xmax, ymin: ymax]
mhd_image[mask_image != 1] = 0
mask_images.append(mask_image)
mhd_images.append(mhd_image)
# show_image(mhd_image)
mask_images = convert2depthlaster(mask_images)
mhd_images = convert2depthlaster(mhd_images)
count += 1
[width, height, depth] = list(np.shape(mhd_images))
patch_count = 1
if width * height >= 900:
patch_step = int(math.sqrt(width * height / 100))
for i in range(patch_size / 2, width - patch_size / 2, patch_step):
for j in range(patch_size / 2, height - patch_size / 2, patch_step):
cur_patch = mhd_images[i - patch_size / 2:i + patch_size / 2,
j - patch_size / 2: j + patch_size / 2, :]
if (np.sum(mask_images[i - patch_size / 2:i + patch_size / 2,
j - patch_size / 2: j + patch_size / 2, :]) / (
(patch_size - 1) * (patch_size - 1) * 3)) < 0.9:
continue
save_path = os.path.join(save_dir, name + '_' + str(patch_count) + '.npy')
# print save_path
np.save(save_path, np.array(cur_patch))
patch_count += 1
if patch_count == 1:
continue
# save_path = os.path.join(save_dir, name + '_' + str(patch_count) + '.png')
# roi_image = Image.fromarray(np.asarray(mhd_images, np.uint8))
# roi_image.save(save_path)
print count
def load_mask_image(image_path, mask_dir_name=['LiverMask', 'TumorMask']):
mask_image = {}
liver_mask = {}
tumors_mask = []
tumor_files_num = len(os.listdir(os.path.join(image_path, mask_dir_name[1]))) / 6
for phase in Config.QIXIANGS:
mask_file_name = glob.glob(os.path.join(image_path, mask_dir_name[0]) + '\\*'+phase+'.mhd')[0]
liver_mask[phase] = read_mhd_image(mask_file_name)
for i in range(tumor_files_num):
tumor_mask_image = {}
for phase in Config.QIXIANGS:
tumor_mask_file_path = \
glob.glob(os.path.join(image_path, mask_dir_name[1]) + '\\*' + phase + '_' + str(i + 1) + '.mhd')[0]
tumor_mask_image[phase] = np.asarray(read_mhd_image(tumor_mask_file_path), np.uint8)
tumors_mask.append(tumor_mask_image)
mask_image['LiverMask'] = liver_mask
mask_image['TumorMask'] = tumors_mask
return mask_image
def extract_ROI_image(mhd_path, tumor_path, liver_path, image_size):
mhd_image = read_mhd_image(mhd_path, rejust=True)
# show_image(mhd_image)
tumor_mask_image = read_mhd_image(tumor_path)
liver_mask_image = read_mhd_image(liver_path)
mhd_image = compress22dim(mhd_image)
tumor_mask_image = compress22dim(tumor_mask_image)
liver_mask_image = compress22dim(liver_mask_image)
# 注意记得填充图像
tumor_bounding_box = get_boundingbox(tumor_mask_image)
liver_average = cal_liver_average(mhd_image, liver_mask_image)
mhd_image_copy = copy(mhd_image)
mhd_image_copy[tumor_mask_image != 1] = 0
roi_image = mhd_image_copy[tumor_bounding_box[0]:tumor_bounding_box[1],
tumor_bounding_box[2]:tumor_bounding_box[3]]
roi_image = resize_image(roi_image, size=image_size)
roi_image = np.asarray(roi_image, np.float32)
roi_image = roi_image / liver_average
return roi_image
def visulization(data_dir):
img = None
for phasename in ['NC', 'ART', 'PV']:
mhd_path = glob(os.path.join(data_dir, phasename + '_Image*.mhd'))[0]
mask_path = glob(
os.path.join(data_dir, phasename + '_Registration*.mhd'))[0]
mhd_image = read_mhd_image(mhd_path)
mask_image = read_mhd_image(mask_path)
mhd_image = np.squeeze(mhd_image)
mask_image = np.squeeze(mask_image)
x_min, x_max, y_min, y_max = get_boundingbox(mask_image)
ROI = mhd_image[x_min:x_max, y_min:y_max]
print np.shape(ROI)
if img is None:
img = []
img.append(ROI)
img = convert2depthlaster(img)
print np.shape(img)
img = Image.fromarray(np.asarray(img, np.uint8))
img.show()
img.save('./multi-phase_ROI.png')
def extract_liver_density(dir_name, suffix_name, save_dir, subclass, type):
'''
提取肝脏的平均密度
:param dir_name: 目前所有病例的存储路径
:param suffix_name: 指定的病灶类型的后缀，比如说cyst 就是０
:param save_dir: 提取得到的ｐａｔｃｈ的存储路径
:return: None
'''
count = 0
names = os.listdir(dir_name)
liver_density_dict = {}
for name in names:
if name.endswith(suffix_name):
# 只提取指定类型病灶的ｐａｔｃｈ
liver_density = []
for phasename in phasenames:
image_path = glob(os.path.join(dir_name, name, phasename + '_Image*.mhd'))[0]
liver_mask_path = glob(os.path.join(dir_name, name, phasename + '*Liver*.mhd'))[0]
mask_path = os.path.join(dir_name, name, phasename + '_Registration.mhd')
mhd_image = read_mhd_image(image_path, rejust=False)
mhd_image = np.squeeze(mhd_image)
# show_image(mhd_image)
mask_image = read_mhd_image(mask_path)
mask_image = np.squeeze(mask_image)
liver_mask_image = read_mhd_image(liver_mask_path)
liver_mask_image = np.squeeze(liver_mask_image)
# 计算肝脏的平均密度
value_sum = np.sum(mhd_image[liver_mask_image != 0])
pixel_num = np.sum(liver_mask_image != 0)
average_value = (1.0 * value_sum) / (1.0 * pixel_num)
liver_density.append(average_value)
print name, liver_density
liver_density_dict[name] = liver_density
scio.savemat(os.path.join(save_dir, 'raw_liver_density_' + str(subclass) + '_' + str(type) + '.mat'), liver_density_dict)
return liver_density_dict
def load_image(image_path):
images = {}
dirs = os.listdir(image_path)
exists = False
for dir_name in dirs:
if dir_name in ['ART', 'NC', 'PV']:
exists = True
mhd_path = os.path.join(image_path, dir_name, dir_name+'.mhd')
if os.path.exists(mhd_path):
images[dir_name] = read_mhd_image(mhd_path)
else:
images[dir_name] = read_dicom_series(os.path.join(image_path, dir_name))
if Config.ADJUST_WW_WC:
images[dir_name] = rejust_pixel_value(images[dir_name])
if not exists:
print image_path, 'happened error'
return images
def load_patch(patch_path, return_roi=False, parent_dir=None):
if not return_roi:
if patch_path.endswith('.jpg'):
return Image.open(patch_path)
if patch_path.endswith('.npy'):
return np.load(patch_path)
else:
phasenames = ['NC', 'ART', 'PV']
if patch_path.endswith('.jpg'):
basename = os.path.basename(patch_path)
basename = basename[:basename.rfind('_')]
mask_images = []
mhd_images = []
for phasename in phasenames:
image_path = glob(
os.path.join(parent_dir, basename,
phasename + '_Image*.mhd'))[0]
mask_path = os.path.join(parent_dir, basename,
phasename + '_Registration.mhd')
mhd_image = read_mhd_image(image_path, rejust=True)
mhd_image = np.squeeze(mhd_image)
# show_image(mhd_image)
mask_image = read_mhd_image(mask_path)
mask_image = np.squeeze(mask_image)
[xmin, xmax, ymin, ymax] = get_boundingbox(mask_image)
# xmin -= 15
# xmax += 15
# ymin -= 15
# ymax += 15
mask_image = mask_image[xmin:xmax, ymin:ymax]
mhd_image = mhd_image[xmin:xmax, ymin:ymax]
mhd_image[mask_image != 1] = 0
mask_images.append(mask_image)
mhd_images.append(mhd_image)
mhd_images = convert2depthlaster(mhd_images)
return mhd_images
if patch_path.endswith('.npy'):
basename = os.path.basename(patch_path)
basename = basename[:basename.rfind('_')]
mask_images = []
mhd_images = []
for phasename in phasenames:
image_path = glob(
os.path.join(parent_dir, basename,
phasename + '_Image*.mhd'))[0]
mask_path = os.path.join(parent_dir, basename,
phasename + '_Registration.mhd')
mhd_image = read_mhd_image(
image_path, rejust=False) # 因为存储的是ｎｐｙ格式，所以不进行窗宽窗位的调整
mhd_image = np.squeeze(mhd_image)
# show_image(mhd_image)
mask_image = read_mhd_image(mask_path)
mask_image = np.squeeze(mask_image)
[xmin, xmax, ymin, ymax] = get_boundingbox(mask_image)
# xmin -= 15
# xmax += 15
# ymin -= 15
# ymax += 15
mask_image = mask_image[xmin:xmax, ymin:ymax]
mhd_image = mhd_image[xmin:xmax, ymin:ymax]
mhd_image[mask_image != 1] = 0
mask_images.append(mask_image)
mhd_images.append(mhd_image)
mhd_images = convert2depthlaster(mhd_images)
return mhd_images
def copy_mhd_image(mhd_path, save_path):
mhd_image = read_mhd_image(mhd_path)
save_mhd_image(mhd_image, save_path)
feed_dict={
x: validation_images,
y_: validation_labels
}
)
_, _, _, error_indexs, error_record = calculate_acc_error(
logits=np.argmax(logits, 1),
label=validation_labels,
show=True
)
print 'accuracy is %g' % \
(validation_accuracy)
return error_indexs, error_record
if __name__ == '__main__':
# va_size, image_w, image_h, channal
dataset = ValDataSet(data_path='/home/give/Documents/dataset/MedicalImage/MedicalImage/ROI/val',
phase='ART',
new_size=[sub_Config.IMAGE_W, sub_Config.IMAGE_H], shuffle=False)
for one_sample_name in dataset.image_names:
original_image = read_mhd_image(one_sample_name)
operations = [
[96, 64, 1.0],
[64, 64, 1.0],
[64, 64, 0.8]
]
error_indexs, error_record = val(
dataset,
load_model_path='/home/give/PycharmProjects/MedicalImage/Net/BaseNet/LeNet/model_finetuing/model_art/',
)
dataset.show_error_name(error_indexs, error_record)
from Tools import read_mhd_image, show_image, save_image_with_mask, compress22dim
import os
from glob import glob
import numpy as np
data_dir = '/home/give/Documents/dataset/MedicalImage/MedicalImage/SL_TrainAndVal/train'
names = os.listdir(data_dir)
for name in names:
cur_dir = os.path.join(data_dir, name)
for phasename in ['NC', 'ART', 'PV']:
image_path = glob(os.path.join(cur_dir, phasename + '_Image*.mhd'))[0]
mask_paths = glob(os.path.join(cur_dir, phasename + '_Mask.mhd'))
if len(mask_paths) == 0:
mask_paths = glob(os.path.join(cur_dir, phasename + '_Mask_*.mhd'))
for path in mask_paths:
if path.endswith('Expand.mhd'):
continue
else:
mask_path = path
break
else:
mask_path = mask_paths[0]
mhd_image = read_mhd_image(image_path, rejust=True)
mhd_image = compress22dim(mhd_image)
print mask_path
tumor_mask_image = read_mhd_image(mask_path)
tumor_mask_image = compress22dim(tumor_mask_image)
print np.max(tumor_mask_image)
save_image_with_mask(mhd_image, tumor_mask_image, None)
input('input %d')
