Datasets:

ArXiv:
DOI:
License:
evsizikova commited on
Commit
430d827
·
verified ·
1 Parent(s): 2f2c487

Upload ssynth_data.py with huggingface_hub

Browse files
Files changed (1) hide show
  1. ssynth_data.py +172 -0
ssynth_data.py ADDED
@@ -0,0 +1,172 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright 2022 for msynth dataset
2
+ #
3
+ # Licensed under the Apache License, Version 2.0 (the "License");
4
+ # you may not use this file except in compliance with the License.
5
+ # You may obtain a copy of the License at
6
+ #
7
+ # http://www.apache.org/licenses/LICENSE-2.0
8
+ #
9
+ # Unless required by applicable law or agreed to in writing, software
10
+ # distributed under the License is distributed on an "AS IS" BASIS,
11
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12
+ # See the License for the specific language governing permissions and
13
+ # limitations under the License.
14
+ '''
15
+ Custom dataset-builder for ssynth dataset
16
+ '''
17
+
18
+ import os
19
+ import datasets
20
+ import glob
21
+ import re
22
+
23
+
24
+ logger = datasets.logging.get_logger(__name__)
25
+
26
+ _CITATION = """\
27
+ @article{kim2024ssynth,
28
+ title={Knowledge-based in silico models and dataset for the comparative evaluation of mammography AI for a range of breast characteristics, lesion conspicuities and doses},
29
+ author={Kim, Andrea and Saharkhiz, Niloufar and Sizikova, Elena and Lago, Miguel, and Sahiner, Berkman and Delfino, Jana G., and Badano, Aldo},
30
+ journal={International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI)},
31
+ volume={},
32
+ pages={},
33
+ year={2024}
34
+ }
35
+ """
36
+
37
+
38
+ _DESCRIPTION = """\
39
+ S-SYNTH is an open-source, flexible skin simulation framework to rapidly generate synthetic skin models and images using digital rendering of an anatomically inspired multi-layer, multi-component skin and growing lesion model. It allows for generation of highly-detailed 3D skin models and digitally rendered synthetic images of diverse human skin tones, with full control of underlying parameters and the image formation process.
40
+ Curated by: Andrea Kim, Niloufar Saharkhiz, Elena Sizikova, Miguel Lago, Berkman Sahiner, Jana Delfino, Aldo Badano
41
+ License: Creative Commons 1.0 Universal License (CC0)
42
+ """
43
+
44
+
45
+ _HOMEPAGE = "https://github.com/DIDSR/ssynth-release?tab=readme-ov-file"
46
+
47
+ _REPO = "https://huggingface.co/datasets/didsr/ssynth_data/resolve/main"
48
+
49
+ # Initialize an empty list to store the file paths
50
+ _CROPPED = True
51
+
52
+ _URLS = {
53
+ "synthetic_data": f"{_REPO}/data/synthetic_dataset/output_10k.zip",
54
+ "read_me": f"{_REPO}/README.md"
55
+ }
56
+
57
+ DATA_DIR = {"all_data": "output_10k"}
58
+
59
+ class ssynth_dataConfig(datasets.BuilderConfig):
60
+ """ssynth dataset"""
61
+ def __init__(self, name, **kwargs):
62
+ super(ssynth_dataConfig, self).__init__(
63
+ version=datasets.Version("1.0.0"),
64
+ name=name,
65
+ description="ssynth_data",
66
+ **kwargs,
67
+ )
68
+
69
+ class ssynth_data(datasets.GeneratorBasedBuilder):
70
+ """ssynth dataset."""
71
+
72
+ DEFAULT_WRITER_BATCH_SIZE = 256
73
+ BUILDER_CONFIGS = [
74
+ ssynth_dataConfig("output_10k"),
75
+ ]
76
+
77
+ def _info(self):
78
+ if self.config.name == "output_10k":
79
+ # Define dataset features and keys
80
+ features = datasets.Features(
81
+ {
82
+ "Cropped": datasets.Features({
83
+ "image": datasets.Value("string"),
84
+ "mask": datasets.Value("string")
85
+ }),
86
+ "Uncropped": datasets.Features({
87
+ "image": datasets.Value("string"),
88
+ "mask": datasets.Value("string")
89
+ })
90
+ }
91
+ )
92
+
93
+ return datasets.DatasetInfo(
94
+ description=_DESCRIPTION,
95
+ features=features,
96
+ supervised_keys=None,
97
+ homepage=_HOMEPAGE,
98
+ citation=_CITATION,
99
+ )
100
+
101
+ def _split_generators(
102
+ self, dl_manager: datasets.utils.download_manager.DownloadManager):
103
+
104
+ if self.config.name == "output_10k":
105
+ data_dir = dl_manager.download_and_extract(_URLS['synthetic_data'])
106
+ return [
107
+ datasets.SplitGenerator(
108
+ name="output_10k",
109
+ gen_kwargs={
110
+ "files": data_dir,
111
+ "name": "all_data",
112
+ },
113
+ ),
114
+ ]
115
+
116
+ def get_all_file_paths(self, root_directory):
117
+ file_paths = [] # List to store file paths
118
+
119
+ # Walk through the directory and its subdirectories using os.walk
120
+ for folder, _, files in os.walk(root_directory):
121
+ for file in files:
122
+ if file == "cropped_image.png":
123
+ # Get the full path of the file
124
+ file_path = os.path.join(folder, file)
125
+ file_paths.append(file_path)
126
+ return file_paths
127
+
128
+ def get_other_images(self, cropped_image_path, file_name):
129
+ other_image_paths = []
130
+
131
+ # Get the directory containing the cropped_image.png
132
+ directory = os.path.dirname(cropped_image_path)
133
+
134
+ # Walk through the directory to find other image files
135
+ for file in os.listdir(directory):
136
+ if file == file_name:
137
+ # Get the full path of the other image file
138
+ file_path = os.path.join(directory, file)
139
+ #other_image_paths.append(file_path)
140
+ return file_path
141
+ return None
142
+
143
+
144
+ def _generate_examples(self, files, name):
145
+ if self.config.name == "output_10k":
146
+ key = 0
147
+ data_paths = self.get_all_file_paths(os.path.join(files, DATA_DIR[name]))
148
+
149
+ cropped_images = []
150
+ uncropped_images = []
151
+ for path in data_paths:
152
+ res_dic = {}
153
+ cropped_image = path
154
+ cropped_mask = self.get_other_images(path,"cropped_mask.png")
155
+ image = self.get_other_images(path,"image.png")
156
+ mask = self.get_other_images(path,"mask.png")
157
+ cropped_data = {
158
+ "image": cropped_image,
159
+ "mask": cropped_mask
160
+ }
161
+ uncropped_data = {
162
+ "image": image,
163
+ "mask": mask
164
+ }
165
+ res_dic["Cropped"] = cropped_data
166
+ res_dic["Uncropped"] = uncropped_data
167
+
168
+ yield key, res_dic
169
+ key += 1
170
+
171
+
172
+