Angelou0516 commited on
Commit
aa035a1
·
verified ·
1 Parent(s): bff35a1

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. README.md +106 -0
  2. dataset_info.json +34 -0
  3. test.parquet +3 -0
  4. train.parquet +3 -0
  5. validation.parquet +3 -0
README.md ADDED
@@ -0,0 +1,106 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ dataset_info:
3
+ features:
4
+ - name: image_id
5
+ dtype: string
6
+ - name: image
7
+ dtype: image
8
+ - name: filename
9
+ dtype: string
10
+ - name: width
11
+ dtype: int32
12
+ - name: height
13
+ dtype: int32
14
+ - name: split
15
+ dtype: string
16
+ - name: label_tumor
17
+ dtype: int8
18
+ - name: label_stroma
19
+ dtype: int8
20
+ - name: label_normal
21
+ dtype: int8
22
+ - name: mask
23
+ dtype: image
24
+ - name: background_mask
25
+ dtype: image
26
+ splits:
27
+ - name: train
28
+ num_examples: 10091
29
+ - name: validation
30
+ num_examples: 40
31
+ - name: test
32
+ num_examples: 80
33
+ license: cc-by-4.0
34
+ task_categories:
35
+ - image-segmentation
36
+ tags:
37
+ - medical
38
+ - pathology
39
+ - lung-cancer
40
+ ---
41
+
42
+ # WSSS4LUAD Dataset
43
+
44
+ This is the WSSS4LUAD dataset for weakly-supervised tissue semantic segmentation in lung adenocarcinoma.
45
+
46
+ ## Dataset Description
47
+
48
+ The dataset contains histopathology image patches from lung adenocarcinoma tissue samples, designed for weakly-supervised semantic segmentation tasks.
49
+
50
+ ## Dataset Structure
51
+
52
+ - **Training**: 10,091 patches with image-level labels
53
+ - **Validation**: 40 patches with pixel-level masks
54
+ - **Test**: 80 patches with pixel-level masks
55
+
56
+ ## Classes
57
+
58
+ The dataset includes 3 tissue classes:
59
+ 1. **Tumor epithelial tissue** (label_tumor)
60
+ 2. **Tumor-associated stroma tissue** (label_stroma)
61
+ 3. **Normal tissue** (label_normal)
62
+
63
+ ## Features
64
+
65
+ - `image_id`: Unique identifier for each image
66
+ - `image`: The histopathology image patch (PNG format, embedded)
67
+ - `filename`: Original filename
68
+ - `width`: Image width in pixels
69
+ - `height`: Image height in pixels
70
+ - `split`: Dataset split (train/validation/test)
71
+ - `label_tumor`: Binary label for tumor epithelial tissue (training only)
72
+ - `label_stroma`: Binary label for tumor stroma tissue (training only)
73
+ - `label_normal`: Binary label for normal tissue (training only)
74
+ - `mask`: Pixel-level segmentation mask (validation/test only)
75
+ - `background_mask`: Background region mask (validation/test only)
76
+
77
+ ## Usage
78
+
79
+ ```python
80
+ from datasets import load_dataset
81
+
82
+ # Load the dataset
83
+ dataset = load_dataset("USERNAME/WSSS4LUAD")
84
+
85
+ # Access different splits
86
+ train_data = dataset['train']
87
+ val_data = dataset['validation']
88
+ test_data = dataset['test']
89
+
90
+ # Example: Load and display an image
91
+ example = train_data[0]
92
+ image = example['image'] # PIL Image
93
+ labels = {
94
+ 'tumor': example['label_tumor'],
95
+ 'stroma': example['label_stroma'],
96
+ 'normal': example['label_normal']
97
+ }
98
+ ```
99
+
100
+ ## Citation
101
+
102
+ If you use this dataset, please cite the original WSSS4LUAD paper.
103
+
104
+ ## License
105
+
106
+ This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.
dataset_info.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset_name": "WSSS4LUAD",
3
+ "description": "Weakly-Supervised Tissue Semantic Segmentation for Lung Adenocarcinoma",
4
+ "num_classes": 3,
5
+ "classes": [
6
+ "tumor_epithelial",
7
+ "tumor_stroma",
8
+ "normal_tissue"
9
+ ],
10
+ "total_samples": 10211,
11
+ "splits": {
12
+ "train": {
13
+ "num_examples": 10091,
14
+ "file_size_mb": 1171.9069919586182,
15
+ "max_row_group_mb": 1.5423202514648438,
16
+ "num_row_groups": 1010
17
+ },
18
+ "validation": {
19
+ "num_examples": 40,
20
+ "file_size_mb": 150.56257247924805,
21
+ "max_row_group_mb": 142.52832126617432,
22
+ "num_row_groups": 4
23
+ },
24
+ "test": {
25
+ "num_examples": 80,
26
+ "file_size_mb": 234.02422714233398,
27
+ "max_row_group_mb": 156.98893070220947,
28
+ "num_row_groups": 8
29
+ }
30
+ },
31
+ "image_format": "PNG (embedded as bytes)",
32
+ "license": "CC BY 4.0",
33
+ "batch_size_used": 10
34
+ }
test.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:931a4ffd9407c8496a752d9276c3300a36c12927a5b53a818bec0405c182697d
3
+ size 245392188
train.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc6fffda4d514d534d0651f9bb7a372eb2e8947235f1d675eb892e8768f8bad2
3
+ size 1228833546
validation.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ebd0e6ea673a91caebbe796afc8ab347e48a3f001eaa122a3a75e4367574b1a
3
+ size 157876300