Datasets:

Formats:
arrow
ArXiv:
Libraries:
Datasets
SushantGautam commited on
Commit
85699d6
·
verified ·
1 Parent(s): f707724

Upload folder using huggingface_hub

Browse files
Files changed (1) hide show
  1. kvasir-points_datasets_script.py +143 -0
kvasir-points_datasets_script.py ADDED
@@ -0,0 +1,143 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
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
+ """Dataset for filtered Kvasir-instrument and Hyper-Kvasir with bounding boxes."""
16
+
17
+ import os
18
+ import json
19
+ from PIL import Image
20
+ import datasets
21
+
22
+ import os
23
+ import json
24
+ import pandas as pd
25
+ import hashlib
26
+
27
+
28
+ cal_mid = lambda bx: [[[float(box['xmin'] + box['xmax']) / 2, float(box['ymin'] + box['ymax']) / 2] for box in bx]]
29
+
30
+
31
+ def cal_sha256(file_path): return hashlib.sha256(
32
+ open(file_path, 'rb').read()).hexdigest()
33
+
34
+
35
+ hyper_label_img_path = '/global/D1/projects/HOST/Datasets/hyper-kvasir/labeled-images/image-labels.csv'
36
+
37
+ hyper_df = pd.read_csv(hyper_label_img_path)
38
+
39
+ hyper_seg_img_path = '/global/D1/projects/HOST/Datasets/hyper-kvasir/segmented-images/bounding-boxes.json'
40
+ hyper_seg_img_base_path = "/global/D1/projects/HOST/Datasets/hyper-kvasir/segmented-images/images"
41
+
42
+ instr_seg_img_path = '/global/D1/projects/HOST/Datasets/kvasir-instrument/bboxes.json'
43
+ instr_seg_img_base_path = '/global/D1/projects/HOST/Datasets/kvasir-instrument/images/'
44
+
45
+ hyper_seg_imgs = json.load(open(hyper_seg_img_path))
46
+ instr_seg_imgs = json.load(open(instr_seg_img_path))
47
+
48
+ _CITATION = """\
49
+ @article{kvasir,
50
+ title={Kvasir-instrument and Hyper-Kvasir datasets for bounding box annotations},
51
+ author={Sushant Gautam and collaborators},
52
+ year={2024}
53
+ }
54
+ """
55
+
56
+ _DESCRIPTION = """
57
+ Filtered Kvasir-instrument and Hyper-Kvasir datasets with bounding boxes for medical imaging tasks.
58
+ Each entry contains images, bounding box coordinates, and additional metadata.
59
+ """
60
+
61
+ _HOMEPAGE = "https://example.com/kvasir-hyper-bbox"
62
+
63
+ _LICENSE = "CC BY-NC 4.0"
64
+
65
+ _URLS = {
66
+ "filtered_data": "https://example.com/kvasir-hyper-bbox-dataset.zip"
67
+ }
68
+
69
+
70
+ class KvasirHyperBBox(datasets.GeneratorBasedBuilder):
71
+ """Dataset for Kvasir-instrument and Hyper-Kvasir with bounding boxes."""
72
+
73
+ VERSION = datasets.Version("1.0.0")
74
+
75
+ BUILDER_CONFIGS = [
76
+ datasets.BuilderConfig(
77
+ name="bbox_dataset",
78
+ version=VERSION,
79
+ description="Dataset with bounding box annotations."
80
+ )
81
+ ]
82
+
83
+ DEFAULT_CONFIG_NAME = "bbox_dataset"
84
+
85
+ def _info(self):
86
+ features = datasets.Features({
87
+ "image_data": datasets.Image(),
88
+ "image_sha256": datasets.Value("string"),
89
+ "points": datasets.Sequence(datasets.Sequence(datasets.Sequence(datasets.Value("float32")))),
90
+ "count": datasets.Value("int64"),
91
+ "label": datasets.Value("string"),
92
+ "collection_method": datasets.Value("string"),
93
+ "classification": datasets.Value("string"),
94
+ "organ": datasets.Value("string")
95
+ })
96
+
97
+ return datasets.DatasetInfo(
98
+ description=_DESCRIPTION,
99
+ homepage=_HOMEPAGE,
100
+ license=_LICENSE,
101
+ citation=_CITATION,
102
+ features=features
103
+ )
104
+
105
+ def _split_generators(self, dl_manager):
106
+ return [
107
+ datasets.SplitGenerator(
108
+ name=datasets.Split.TRAIN,
109
+ gen_kwargs={},
110
+ )
111
+ ]
112
+
113
+ def _generate_examples(self):
114
+ for key, entry in hyper_seg_imgs.items():
115
+ img_path = os.path.join(hyper_seg_img_base_path, f"{key}.jpg")
116
+ hyper_entry = hyper_df.loc[hyper_df['Video file'] == key].iloc[0]
117
+ yield key, {
118
+ "image_data": open(img_path, 'rb').read(),
119
+ "image_sha256": cal_sha256(img_path),
120
+ "points": cal_mid(entry['bbox']),
121
+ "count": len(entry['bbox']),
122
+ "label": hyper_entry.Finding,
123
+ "collection_method": 'counting',
124
+ "classification": hyper_entry.Classification,
125
+ "organ": hyper_entry.Organ
126
+ }
127
+
128
+ for key, entry in instr_seg_imgs.items():
129
+ img_path = os.path.join(instr_seg_img_base_path, f"{key}.jpg")
130
+ assert len(cal_mid(entry['bbox'])) > 0
131
+ yield key, {
132
+ "image_data": open(img_path, 'rb').read(),
133
+ "image_sha256": cal_sha256(img_path),
134
+ "points": cal_mid(entry['bbox']),
135
+ "count": len(entry['bbox']),
136
+ "label": "instrument",
137
+ "collection_method": "counting",
138
+ "classification": "instrument",
139
+ "organ": "instrument"
140
+ }
141
+
142
+ #datasets-cli test /global/D1/projects/HOST/Datasets/hyper-kvasir/sushant-experiments/kvasir-points_datasets_script.py --save_info --all_configs --trust_remote_code
143
+ # huggingface-cli upload kvasir-points . . --repo-type dataset