README: 14 datasets
Browse files
README.md
CHANGED
|
@@ -94,6 +94,18 @@ configs:
|
|
| 94 |
data_files:
|
| 95 |
- split: train
|
| 96 |
path: sidecar/Oculo/*.parquet
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 97 |
- config_name: RAVIR
|
| 98 |
data_files:
|
| 99 |
- split: train
|
|
@@ -166,7 +178,7 @@ configs:
|
|
| 166 |
|
| 167 |
# OphGen — curated public ophthalmic datasets
|
| 168 |
|
| 169 |
-
Curated copies of
|
| 170 |
images as stored in the OphGen image store (de-identified, de-duplicated, native resolution), the train / validation / test
|
| 171 |
split used in the project, six captions per image (short / medium / long × English / Chinese) and all original task annotations.
|
| 172 |
|
|
@@ -194,13 +206,14 @@ In `all`, an image released by several datasets appears once per dataset (same `
|
|
| 194 |
| [GLEAM](cards/GLEAM.md) | GLEAM | IR/NIR, OCT | 1,436 | 480 | 480 | 1 |
|
| 195 |
| [IDRiD](cards/IDRiD.md) | IDRiD (segmentation and grading parts) | CFP | 382 | 42 | 107 | 2 |
|
| 196 |
| [Oculo](cards/Oculo.md) | Oculo | Conventional ocular B-scan ultrasonography | 1,189 | 162 | 245 | 2 |
|
|
|
|
| 197 |
| [RAVIR](cards/RAVIR.md) | RAVIR | IR/NIR | 20 | 2 | 19 | 2 |
|
| 198 |
| [REFUGE2](cards/REFUGE2.md) | REFUGE2 | CFP | 395 | 402 | 403 | 3 |
|
| 199 |
| [ROSSA](cards/ROSSA.md) | ROSSA | OCTA | 100 | 100 | 100 | 1 |
|
| 200 |
| [RVO-ME](cards/RVO-ME.md) | RVO-ME | OCT | 2,057 | 229 | 580 | 3 |
|
| 201 |
| [SJTU_Peripapillary](cards/SJTU_Peripapillary.md) | SJTU peripapillary OCT | OCT | 74 | 24 | 24 | 3 |
|
| 202 |
| [UWF-DR_2026](cards/UWF-DR_2026.md) | UWF-DR | UWF | 1,135 | 158 | 327 | 2 |
|
| 203 |
-
| **Total** | | | 9,
|
| 204 |
|
| 205 |
Split rules: `official` = official split; `official_val_from_train` = 10% of the official training set (by patient) used as
|
| 206 |
validation; `labeled_7_1_2` = patient-level 7:1:2 split with a fixed seed (no official split); `pretrain_90_5_5` = 90:5:5 split
|
|
|
|
| 94 |
data_files:
|
| 95 |
- split: train
|
| 96 |
path: sidecar/Oculo/*.parquet
|
| 97 |
+
- config_name: ORIGA
|
| 98 |
+
data_files:
|
| 99 |
+
- split: train
|
| 100 |
+
path: data/ORIGA/train-*.parquet
|
| 101 |
+
- split: validation
|
| 102 |
+
path: data/ORIGA/validation-*.parquet
|
| 103 |
+
- split: test
|
| 104 |
+
path: data/ORIGA/test-*.parquet
|
| 105 |
+
- config_name: ORIGA-sidecar
|
| 106 |
+
data_files:
|
| 107 |
+
- split: train
|
| 108 |
+
path: sidecar/ORIGA/*.parquet
|
| 109 |
- config_name: RAVIR
|
| 110 |
data_files:
|
| 111 |
- split: train
|
|
|
|
| 178 |
|
| 179 |
# OphGen — curated public ophthalmic datasets
|
| 180 |
|
| 181 |
+
Curated copies of 14 public ophthalmic imaging datasets, prepared for the OphGen ophthalmic generative-model project:
|
| 182 |
images as stored in the OphGen image store (de-identified, de-duplicated, native resolution), the train / validation / test
|
| 183 |
split used in the project, six captions per image (short / medium / long × English / Chinese) and all original task annotations.
|
| 184 |
|
|
|
|
| 206 |
| [GLEAM](cards/GLEAM.md) | GLEAM | IR/NIR, OCT | 1,436 | 480 | 480 | 1 |
|
| 207 |
| [IDRiD](cards/IDRiD.md) | IDRiD (segmentation and grading parts) | CFP | 382 | 42 | 107 | 2 |
|
| 208 |
| [Oculo](cards/Oculo.md) | Oculo | Conventional ocular B-scan ultrasonography | 1,189 | 162 | 245 | 2 |
|
| 209 |
+
| [ORIGA](cards/ORIGA.md) | ORIGA | CFP | 293 | 32 | 325 | 4 |
|
| 210 |
| [RAVIR](cards/RAVIR.md) | RAVIR | IR/NIR | 20 | 2 | 19 | 2 |
|
| 211 |
| [REFUGE2](cards/REFUGE2.md) | REFUGE2 | CFP | 395 | 402 | 403 | 3 |
|
| 212 |
| [ROSSA](cards/ROSSA.md) | ROSSA | OCTA | 100 | 100 | 100 | 1 |
|
| 213 |
| [RVO-ME](cards/RVO-ME.md) | RVO-ME | OCT | 2,057 | 229 | 580 | 3 |
|
| 214 |
| [SJTU_Peripapillary](cards/SJTU_Peripapillary.md) | SJTU peripapillary OCT | OCT | 74 | 24 | 24 | 3 |
|
| 215 |
| [UWF-DR_2026](cards/UWF-DR_2026.md) | UWF-DR | UWF | 1,135 | 158 | 327 | 2 |
|
| 216 |
+
| **Total** | | | 9,647 | 2,050 | 2,978 | |
|
| 217 |
|
| 218 |
Split rules: `official` = official split; `official_val_from_train` = 10% of the official training set (by patient) used as
|
| 219 |
validation; `labeled_7_1_2` = patient-level 7:1:2 split with a fixed seed (no official split); `pretrain_90_5_5` = 90:5:5 split
|