Hugging Face's logo

Spaces:
andreped
/
LungTumorMask
Sleeping

App Files Community
LungTumorMask
File size: 8,337 Bytes 
1cc0005
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
 
import glob
import platform
import nibabel
import os
import numpy as np
import torch
from monai.transforms import Compose, Spacingd, ToTensord, LoadImaged, AddChanneld, ThresholdIntensityd, LoadImage, ScaleIntensityRanged, NormalizeIntensityd
from monai.data import NiftiSaver
from segment_to_segmentbox import generate_segmentbox
import pickle
from Logger.loggingservice import Logger

logger = Logger("http://82.194.207.154:5000/api/log", "XqnJHdalUd")

directory_split = "\\" if platform.system() == "Windows" else "/"

voxel_normalizer = Compose(
        [
            LoadImaged(keys=['image', 'boxes', 'label']),
            ThresholdIntensityd(keys=['image'], above = False, threshold = 1000, cval = 1000),
            ThresholdIntensityd(keys=['image'], above = True, threshold = -1024, cval = -1024),
            AddChanneld(keys=['image', 'boxes', 'label']),
            Spacingd(keys=['image', 'boxes', 'label'], pixdim=(0.7, 0.7, 0.7)),
            #ScaleIntensityRanged(keys=['image'], a_min=-1024, a_max=400, b_min=-1, b_max=1),
            NormalizeIntensityd(keys=['image']),
            ThresholdIntensityd(keys=['boxes'], above = False, threshold = 0.5, cval = 1),
            ThresholdIntensityd(keys=['boxes'], above = True, threshold = 0.5, cval = 0),
            ToTensord(keys=['image', 'boxes', 'label'])
        ]
    )

# voxel_normalizer = Compose(
#     [
#         LoadImaged(keys=['label']),
#         AddChanneld(keys=['label']),
#         Spacingd(keys=['label'], pixdim=(0.7, 0.7, 0.7)),
#         ToTensord(keys=['label'])
#     ]
# )


def crop(image, center, dimensions):
    #holder = np.zeros(shape=dimensions)
    holder = np.full(dimensions, -1).astype(np.float32)
    
    crop = image[(center[0]-int(dimensions[0]/2)):center[0]+int(dimensions[0]/2), 
        (center[1]-int(dimensions[1]/2)):center[1]+int(dimensions[1]/2), 
        (center[2]-int(dimensions[2]/2)):center[2]+int(dimensions[2]/2)]

    #return crop
    holder[:crop.shape[0], :crop.shape[1], : crop.shape[2]] = crop
    return holder

def find_center(box_array):
    
    z_min = -1
    z_max = 0
    y_min = float('inf')
    y_max = 0
    x_min = float('inf')
    x_max = 0


    box_array = box_array.swapaxes(0, 2)
    for i, layer in enumerate(box_array):
        if(np.amax(layer) < 1):
            continue
        if(z_min == -1):
            z_min = i
        z_max = i

        y = np.any(layer, axis = 1)
        x = np.any(layer, axis = 0)
        y_minl, y_maxl = np.argmax(y) + 1, layer.shape[0] - np.argmax(np.flipud(y))
        x_minl, x_maxl = np.argmax(x) + 1, layer.shape[1] - np.argmax(np.flipud(x))

        if(y_minl < y_min):
            y_min = y_minl
        if(x_minl < x_min):
            x_min = x_minl
        if(y_maxl > y_max):
            y_max = y_maxl
        if(x_maxl > x_max):
            x_max = x_maxl
    
    return x_min + int((x_max - x_min)/2), y_min + int((y_max - y_min)/2), z_min + int((z_max - z_min)/2)

def crop_and_store(root_folder, save_folder):    
    #os.mkdir(f"{save_folder}{directory_split}Images")
    #os.mkdir(f"{save_folder}{directory_split}Boxes")
    #os.mkdir(f"{save_folder}{directory_split}Labels")

    image_directory = f'{root_folder}{directory_split}Images{directory_split}'
    box_directory = f'{root_folder}{directory_split}Boxes{directory_split}'
    label_directory = f'{root_folder}{directory_split}Labels{directory_split}'

    crop_corner_dict = {}

    for box_file in glob.glob(f'{box_directory}*'):
        file_name = box_file.split(directory_split)[-1]
        print(f"Processing {file_name}")

        data = [{'image' : f"{image_directory}{file_name}", 'boxes' : f"{box_directory}{file_name}", 'label': f"{label_directory}{file_name}"}]

        #imloader = LoadImage()
        #org_box_shape = imloader(data[0]['boxes'])[0].shape

        data = voxel_normalizer(data)

        
        box = nibabel.load(f"{box_directory}{file_name}")
        new_box_arr = torch.nn.functional.interpolate(torch.Tensor(box.get_fdata()).unsqueeze(0).unsqueeze(0), 
            data[0]['image'].numpy().shape[1:], mode='trilinear', align_corners = False).squeeze(0).squeeze(0)
        new_box_arr[new_box_arr < 0.5] = 0
        new_box_arr[new_box_arr >= 0.5] = 1
        #new_box_arr =

        #new_box_arr = data[0]['boxes'].squeeze()
        try:
            center = find_center(np.copy(new_box_arr.numpy()))
        except:
            print(f"Couldnt calculate center of {file_name}")
            continue

    
        centered_image = crop(data[0]['image'].squeeze(0).numpy(), center, (128, 128, 128))
        #centered_box = crop(data[0]['boxes'].squeeze(0).numpy(), center, (128, 128, 128))
        centered_box = crop(new_box_arr, center, (128, 128, 128))
        centered_label = crop(data[0]['label'].squeeze(0).numpy(), center, (128, 128, 128))
        centered_label[centered_label >= 0.5] = 1
        centered_label[centered_label < 0.5] = 0
        #centered_box = generate_segmentbox(np.copy(centered_label))
        #centered_box[centered_box < 0.5] = 0
        #centered_box[centered_box > 0.5] = 1


        affine = data[0]['image_meta_dict']['affine']

        box_nifti = nibabel.Nifti1Image(centered_box, affine)
        image_nifti = nibabel.Nifti1Image(centered_image, affine)
        label_nifti = nibabel.Nifti1Image(centered_label, affine)

        nibabel.save(box_nifti, f"{save_folder}{directory_split}Boxes{directory_split}{file_name}")
        nibabel.save(image_nifti, f"{save_folder}{directory_split}Images{directory_split}{file_name}")
        nibabel.save(label_nifti, f"{save_folder}{directory_split}Labels{directory_split}{file_name}")

        org_box_shape = box.get_fdata().shape
        crop_corner_dict[file_name] = (center, data[0]['image'].squeeze(0).numpy().shape, org_box_shape)
        print(f"Saving {file_name}")
    

    pickle.dump(crop_corner_dict, open( f"{save_folder}{directory_split}crop_reference.bin", "wb" ))

def inject_image(holder, image, corner):

    x_max = min(corner[0]+image.shape[0], holder.shape[0])
    y_max = min(corner[1]+image.shape[1], holder.shape[1])
    z_max = min(corner[2]+image.shape[2], holder.shape[2])

    x_imslice = corner[0]+image.shape[0] - x_max
    y_imslice = corner[1]+image.shape[1] - y_max
    z_imslice = corner[2]+image.shape[2] - z_max

    holder[corner[0]:x_max, corner[1]:y_max, corner[2]: z_max] = image[:image.shape[0] - x_imslice, :image.shape[1] - y_imslice, :image.shape[2] - z_imslice]
    return holder

def calculate_corner(imputation_shape, center_in_holder):
    x = center_in_holder[0] - imputation_shape[0]/2
    y = center_in_holder[1] - imputation_shape[1]/2
    z = center_in_holder[2] - imputation_shape[2]/2
    return (int(x), int(y), int(z))

def load_and_relocate(reference_file, segment_folder, store_folder):

    ref_dict = pickle.load(open(reference_file, 'rb'))
    
    logger.LogInfo("Starting load and relocate", [])
    for image_file in glob.glob(f"{segment_folder}{directory_split}*.nii.gz"):
        try:        
            file_name = image_file.split(directory_split)[-1]
            references = ref_dict[file_name[4:]]

            holder = np.zeros(shape=references[1])
            nib = nibabel.load(image_file)
            image = nib.get_fdata()
            corner = calculate_corner(image.shape, references[0])
            injected = inject_image(holder, image, corner)

            injected = torch.nn.functional.interpolate(torch.Tensor(injected).unsqueeze(0).unsqueeze(0), size=references[-1], mode='trilinear').squeeze(0).squeeze(0)
            injected[injected >= 0.5] = 1
            injected[injected < 0.5] = 0
            injected = injected.numpy()
            nifti = nibabel.Nifti1Image(injected, nib.affine)
            nibabel.save(nifti, f"{save_folder}{directory_split}{file_name}")

            print(file_name)
        except:
            logger.LogWarning(f"Failed relocate", [image_file])



if __name__ == "__main__":

    crop_and_store("D:\\Datasets\\Temp\\", "D:\\Datasets\\Temp\\Save\\")

    #ref = "/home/tumor/data/LungDx-Tumor/crop_reference.bin"
    #segment_folder = "/home/tumor/data/LungDx-Tumor/Labels/"
    #save_folder = "/home/tumor/data/LungDx-Full/Labels/"

    #load_and_relocate(ref, segment_folder, save_folder)
    #logger.LogMilestone("Finished relocating", [])