Add files using upload-large-folder tool
Browse files
cards/Duke_DME.md
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Duke_DME
|
| 2 |
+
|
| 3 |
+
Curated copy of the public dataset **Duke DME (Chiu)**. Config `Duke_DME`.
|
| 4 |
+
**Use and licence follow the original dataset.** Please cite the original publication and respect its terms.
|
| 5 |
+
|
| 6 |
+
## Original dataset
|
| 7 |
+
|
| 8 |
+
| Field | Value |
|
| 9 |
+
| :--- | :--- |
|
| 10 |
+
| Country / region | USA |
|
| 11 |
+
| Acquisition centre | Duke University |
|
| 12 |
+
| Device | Heidelberg Spectralis HRA+OCT (as cited by Li et al. 2021) |
|
| 13 |
+
| Population | 10 patients with severe DME, 11 B-scans each (110); this copy is cropped in width |
|
| 14 |
+
| Original annotations | Retinal layer and fluid segmentation (grey-coded masks with two undocumented code sets in this copy) |
|
| 15 |
+
| Reference / link | Chiu SJ, Allingham MJ, Mettu PS, et al. Kernel regression based segmentation of optical coherence tomography images with diabetic macular edema. Biomed Opt Express. 2015;6(4):1172-1194. doi:10.1364/BOE.6.001172; https://people.duke.edu/~sf59/Chiu_BOE_2014_dataset.htm |
|
| 16 |
+
|
| 17 |
+
## Contents
|
| 18 |
+
|
| 19 |
+
- Images: 110 (train 77, validation 11, test 22); modalities: OCT 110.
|
| 20 |
+
- Split rule: labeled_7_1_2 (110).
|
| 21 |
+
- Task labels: 440 rows; raster masks embedded: 110; other annotation files: 0.
|
| 22 |
+
|
| 23 |
+
| Task | Label name | Rows |
|
| 24 |
+
| :--- | :--- | ---: |
|
| 25 |
+
| classification | patient_diagnosis (cohort: DME; patient level) | 110 |
|
| 26 |
+
| grouping | bscan_index | 110 |
|
| 27 |
+
| grouping | patient_id | 110 |
|
| 28 |
+
| segmentation | duke_dme_layers_fluid (grey-coded as received; 2 undocumented code sets) | 110 |
|
data/Duke_DME/test-00000-of-00001.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b4e381a79b8411ffe6920b279d1b6315606846f6061074baf080c3d8067e3926
|
| 3 |
+
size 3567316
|
data/Duke_DME/train-00000-of-00001.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8b2cdf7289c960ea4b1f5bc1f506703b18fa43d2f29c810fdd605cd68240ead
|
| 3 |
+
size 11746147
|
data/Duke_DME/validation-00000-of-00001.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc2406b0f3d12f4f010be8d7c26c28d7e8205485eaf0cb0d19e960ea3981dada
|
| 3 |
+
size 1609282
|
manifests/Duke_DME.json
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset": "Duke_DME",
|
| 3 |
+
"images": {
|
| 4 |
+
"train": 77,
|
| 5 |
+
"validation": 11,
|
| 6 |
+
"test": 22
|
| 7 |
+
},
|
| 8 |
+
"labels": 440,
|
| 9 |
+
"masks_embedded": 110,
|
| 10 |
+
"files": {
|
| 11 |
+
"data/Duke_DME/test-00000-of-00001.parquet": 3567316,
|
| 12 |
+
"data/Duke_DME/train-00000-of-00001.parquet": 11746147,
|
| 13 |
+
"data/Duke_DME/validation-00000-of-00001.parquet": 1609282
|
| 14 |
+
},
|
| 15 |
+
"sidecar_files": 0,
|
| 16 |
+
"counts_by_modality": {
|
| 17 |
+
"oct_bscan": 110
|
| 18 |
+
},
|
| 19 |
+
"split_rules": {
|
| 20 |
+
"labeled_7_1_2": 110
|
| 21 |
+
},
|
| 22 |
+
"tasks": [
|
| 23 |
+
[
|
| 24 |
+
"classification",
|
| 25 |
+
"patient_diagnosis (cohort: DME; patient level)",
|
| 26 |
+
110
|
| 27 |
+
],
|
| 28 |
+
[
|
| 29 |
+
"grouping",
|
| 30 |
+
"bscan_index",
|
| 31 |
+
110
|
| 32 |
+
],
|
| 33 |
+
[
|
| 34 |
+
"grouping",
|
| 35 |
+
"patient_id",
|
| 36 |
+
110
|
| 37 |
+
],
|
| 38 |
+
[
|
| 39 |
+
"segmentation",
|
| 40 |
+
"duke_dme_layers_fluid (grey-coded as received; 2 undocumented code sets)",
|
| 41 |
+
110
|
| 42 |
+
]
|
| 43 |
+
],
|
| 44 |
+
"name": "Duke DME (Chiu)"
|
| 45 |
+
}
|