ziq commited on
Commit
bbf728c
·
1 Parent(s): cf78c15

Update rsna-atd.py

Browse files
Files changed (1) hide show
  1. rsna-atd.py +61 -22
rsna-atd.py CHANGED
@@ -32,26 +32,26 @@ class RSNAATD(datasets.GeneratorBasedBuilder):
32
  description=_DESCRIPTION,
33
  features=datasets.Features(
34
  {
35
- # "patient_id": datasets.Value("int64"),
36
- # "series_id": datasets.Value("int64"),
37
- # "aortic_hu": datasets.Value("float64"),
38
- # "incomplete_organ": datasets.Value("int64"),
39
- # "bowel_healthy": datasets.Value("int64"),
40
- # "bowel_injury": datasets.Value("int64"),
41
- # "extravasation_healthy": datasets.Value("int64"),
42
- # "extravasation_injury": datasets.Value("int64"),
43
- # "kidney_healthy": datasets.Value("int64"),
44
- # "kidney_low": datasets.Value("int64"),
45
- # "kidney_high": datasets.Value("int64"),
46
- # "liver_healthy": datasets.Value("int64"),
47
- # "liver_low": datasets.Value("int64"),
48
- # "liver_high": datasets.Value("int64"),
49
- # "spleen_healthy": datasets.Value("int64"),
50
- # "spleen_low": datasets.Value("int64"),
51
- # "spleen_high": datasets.Value("int64"),
52
- # "any_injury": datasets.Value("int64"),
53
- "image": datasets.Array3D(shape=(None, 512, 512), dtype=np.uint8),
54
- "mask": datasets.Array3D(shape=(None, 512, 512), dtype=np.uint8),
55
  }
56
  ),
57
  supervised_keys=None,
@@ -81,9 +81,48 @@ class RSNAATD(datasets.GeneratorBasedBuilder):
81
  def _generate_examples(self, images, masks, metadata):
82
  df = pd.read_csv(metadata)
83
  for idx, data in enumerate(df.to_numpy()):
84
- image, mask = hkl.load(f"images/{data[0]}"), hkl.load(f"masks/{data[0]}")
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
85
 
86
- yield idx, {"image": image, "mask": mask}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
87
 
88
  # for idx, ((image_path), (mask_path)) in enumerate(zip(images, masks)):
89
  # # row = df.loc[df["path"] == image_path.lower().replace("images/", "")]
 
32
  description=_DESCRIPTION,
33
  features=datasets.Features(
34
  {
35
+ "patient_id": datasets.Value("int64"),
36
+ "series_id": datasets.Value("int64"),
37
+ "aortic_hu": datasets.Value("float64"),
38
+ "incomplete_organ": datasets.Value("int64"),
39
+ "bowel_healthy": datasets.Value("int64"),
40
+ "bowel_injury": datasets.Value("int64"),
41
+ "extravasation_healthy": datasets.Value("int64"),
42
+ "extravasation_injury": datasets.Value("int64"),
43
+ "kidney_healthy": datasets.Value("int64"),
44
+ "kidney_low": datasets.Value("int64"),
45
+ "kidney_high": datasets.Value("int64"),
46
+ "liver_healthy": datasets.Value("int64"),
47
+ "liver_low": datasets.Value("int64"),
48
+ "liver_high": datasets.Value("int64"),
49
+ "spleen_healthy": datasets.Value("int64"),
50
+ "spleen_low": datasets.Value("int64"),
51
+ "spleen_high": datasets.Value("int64"),
52
+ "any_injury": datasets.Value("int64"),
53
+ # "image": datasets.Array3D(shape=(None, 512, 512), dtype=np.uint8),
54
+ # "mask": datasets.Array3D(shape=(None, 512, 512), dtype=np.uint8),
55
  }
56
  ),
57
  supervised_keys=None,
 
81
  def _generate_examples(self, images, masks, metadata):
82
  df = pd.read_csv(metadata)
83
  for idx, data in enumerate(df.to_numpy()):
84
+ # image, mask = hkl.load(f"images/{data[0]}"), hkl.load(f"masks/{data[0]}")
85
+ (
86
+ patient_id,
87
+ series_id,
88
+ aortic_hu,
89
+ incomplete_organ,
90
+ bowel_healthy,
91
+ bowel_injury,
92
+ extravasation_healthy,
93
+ extravasation_injury,
94
+ kidney_healthy,
95
+ kidney_low,
96
+ kidney_high,
97
+ liver_healthy,
98
+ liver_low,
99
+ liver_high,
100
+ spleen_healthy,
101
+ spleen_low,
102
+ spleen_high,
103
+ any_injury,
104
+ ) = data[1:]
105
 
106
+ yield idx, {
107
+ "patient_id": patient_id,
108
+ "series_id": series_id,
109
+ "aortic_hu": aortic_hu,
110
+ "incomplete_organ": incomplete_organ,
111
+ "bowel_healthy": bowel_healthy,
112
+ "bowel_injury": bowel_injury,
113
+ "extravasation_healthy": extravasation_healthy,
114
+ "extravasation_injury": extravasation_injury,
115
+ "kidney_healthy": kidney_healthy,
116
+ "kidney_low": kidney_low,
117
+ "kidney_high": kidney_high,
118
+ "liver_healthy": liver_healthy,
119
+ "liver_low": liver_low,
120
+ "liver_high": liver_high,
121
+ "spleen_healthy": spleen_healthy,
122
+ "spleen_low": spleen_low,
123
+ "spleen_high": spleen_high,
124
+ "any_injury": any_injury,
125
+ }
126
 
127
  # for idx, ((image_path), (mask_path)) in enumerate(zip(images, masks)):
128
  # # row = df.loc[df["path"] == image_path.lower().replace("images/", "")]