README: 8 datasets
Browse files
README.md
CHANGED
|
@@ -24,6 +24,18 @@ configs:
|
|
| 24 |
data_files:
|
| 25 |
- split: train
|
| 26 |
path: sidecar/DRIVE/*.parquet
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 27 |
- config_name: IDRiD
|
| 28 |
data_files:
|
| 29 |
- split: train
|
|
@@ -96,7 +108,7 @@ configs:
|
|
| 96 |
|
| 97 |
# OphGen — curated public ophthalmic datasets
|
| 98 |
|
| 99 |
-
Curated copies of
|
| 100 |
images as stored in the OphGen image store (de-identified, de-duplicated, native resolution), the train / validation / test
|
| 101 |
split used in the project, six captions per image (short / medium / long × English / Chinese) and all original task annotations.
|
| 102 |
|
|
@@ -118,13 +130,14 @@ In `all`, an image released by several datasets appears once per dataset (same `
|
|
| 118 |
| Config | Dataset | Modalities | Train | Validation | Test | Tasks |
|
| 119 |
| :--- | :--- | :--- | ---: | ---: | ---: | ---: |
|
| 120 |
| [DRIVE](cards/DRIVE.md) | DRIVE | CFP | 18 | 2 | 20 | 1 |
|
|
|
|
| 121 |
| [IDRiD](cards/IDRiD.md) | IDRiD (segmentation and grading parts) | CFP | 382 | 42 | 107 | 2 |
|
| 122 |
| [RAVIR](cards/RAVIR.md) | RAVIR | IR/NIR | 20 | 2 | 19 | 2 |
|
| 123 |
| [REFUGE2](cards/REFUGE2.md) | REFUGE2 | CFP | 395 | 402 | 403 | 3 |
|
| 124 |
| [ROSSA](cards/ROSSA.md) | ROSSA | OCTA | 100 | 100 | 100 | 1 |
|
| 125 |
| [RVO-ME](cards/RVO-ME.md) | RVO-ME | OCT | 2,057 | 229 | 580 | 3 |
|
| 126 |
| [SJTU_Peripapillary](cards/SJTU_Peripapillary.md) | SJTU peripapillary OCT | OCT | 74 | 24 | 24 | 3 |
|
| 127 |
-
| **Total** | | |
|
| 128 |
|
| 129 |
Split rules: `official` = official split; `official_val_from_train` = 10% of the official training set (by patient) used as
|
| 130 |
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
|
|
|
|
| 24 |
data_files:
|
| 25 |
- split: train
|
| 26 |
path: sidecar/DRIVE/*.parquet
|
| 27 |
+
- config_name: GLEAM
|
| 28 |
+
data_files:
|
| 29 |
+
- split: train
|
| 30 |
+
path: data/GLEAM/train-*.parquet
|
| 31 |
+
- split: validation
|
| 32 |
+
path: data/GLEAM/validation-*.parquet
|
| 33 |
+
- split: test
|
| 34 |
+
path: data/GLEAM/test-*.parquet
|
| 35 |
+
- config_name: GLEAM-sidecar
|
| 36 |
+
data_files:
|
| 37 |
+
- split: train
|
| 38 |
+
path: sidecar/GLEAM/*.parquet
|
| 39 |
- config_name: IDRiD
|
| 40 |
data_files:
|
| 41 |
- split: train
|
|
|
|
| 108 |
|
| 109 |
# OphGen — curated public ophthalmic datasets
|
| 110 |
|
| 111 |
+
Curated copies of 8 public ophthalmic imaging datasets, prepared for the OphGen ophthalmic generative-model project:
|
| 112 |
images as stored in the OphGen image store (de-identified, de-duplicated, native resolution), the train / validation / test
|
| 113 |
split used in the project, six captions per image (short / medium / long × English / Chinese) and all original task annotations.
|
| 114 |
|
|
|
|
| 130 |
| Config | Dataset | Modalities | Train | Validation | Test | Tasks |
|
| 131 |
| :--- | :--- | :--- | ---: | ---: | ---: | ---: |
|
| 132 |
| [DRIVE](cards/DRIVE.md) | DRIVE | CFP | 18 | 2 | 20 | 1 |
|
| 133 |
+
| [GLEAM](cards/GLEAM.md) | GLEAM | IR/NIR, OCT | 1,436 | 480 | 480 | 1 |
|
| 134 |
| [IDRiD](cards/IDRiD.md) | IDRiD (segmentation and grading parts) | CFP | 382 | 42 | 107 | 2 |
|
| 135 |
| [RAVIR](cards/RAVIR.md) | RAVIR | IR/NIR | 20 | 2 | 19 | 2 |
|
| 136 |
| [REFUGE2](cards/REFUGE2.md) | REFUGE2 | CFP | 395 | 402 | 403 | 3 |
|
| 137 |
| [ROSSA](cards/ROSSA.md) | ROSSA | OCTA | 100 | 100 | 100 | 1 |
|
| 138 |
| [RVO-ME](cards/RVO-ME.md) | RVO-ME | OCT | 2,057 | 229 | 580 | 3 |
|
| 139 |
| [SJTU_Peripapillary](cards/SJTU_Peripapillary.md) | SJTU peripapillary OCT | OCT | 74 | 24 | 24 | 3 |
|
| 140 |
+
| **Total** | | | 4,482 | 1,281 | 1,733 | |
|
| 141 |
|
| 142 |
Split rules: `official` = official split; `official_val_from_train` = 10% of the official training set (by patient) used as
|
| 143 |
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
|