README: 10 datasets
Browse files
README.md
CHANGED
|
@@ -12,6 +12,14 @@ configs:
|
|
| 12 |
path: data/*/validation-*.parquet
|
| 13 |
- split: test
|
| 14 |
path: data/*/test-*.parquet
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
- config_name: chengdu_ffa
|
| 16 |
data_files:
|
| 17 |
- split: train
|
|
@@ -88,7 +96,7 @@ configs:
|
|
| 88 |
|
| 89 |
# OphGen — curated institutional ophthalmic datasets
|
| 90 |
|
| 91 |
-
Curated copies of
|
| 92 |
images as stored in the OphGen image store (de-identified, de-duplicated, native resolution), the train / validation / test
|
| 93 |
split used in the project, six captions per image (short / medium / long × English / Chinese) and all original task annotations.
|
| 94 |
|
|
@@ -98,8 +106,8 @@ split used in the project, six captions per image (short / medium / long × Engl
|
|
| 98 |
|
| 99 |
```python
|
| 100 |
from datasets import load_dataset
|
| 101 |
-
ds = load_dataset("Kaphathy/OphGen-Private", "
|
| 102 |
-
side = load_dataset("Kaphathy/OphGen-Private", "
|
| 103 |
pre = load_dataset("Kaphathy/OphGen-Private", "all") # every dataset (default config), e.g. for pretraining
|
| 104 |
```
|
| 105 |
|
|
@@ -109,6 +117,7 @@ In `all`, an image released by several datasets appears once per dataset (same `
|
|
| 109 |
|
| 110 |
| Config | Dataset | Modalities | Train | Validation | Test | Tasks |
|
| 111 |
| :--- | :--- | :--- | ---: | ---: | ---: | ---: |
|
|
|
|
| 112 |
| [chengdu_ffa](cards/chengdu_ffa.md) | chengdu_ffa | FFA | 3,519 | 473 | 934 | 1 |
|
| 113 |
| [eyFRDCls](cards/eyFRDCls.md) | eyFRDCls | CFP | 418 | 53 | 139 | 1 |
|
| 114 |
| [FQ_Datasets](cards/FQ_Datasets.md) | FQ_Datasets | CFP | 2,681 | 402 | 802 | 1 |
|
|
@@ -118,7 +127,7 @@ In `all`, an image released by several datasets appears once per dataset (same `
|
|
| 118 |
| [trhd3_FRD_Cls66_Train](cards/trhd3_FRD_Cls66_Train.md) | trhd3_FRD_Cls66_Train | CFP | 7,780 | 1,106 | 2,176 | 1 |
|
| 119 |
| [yinhai_ffa](cards/yinhai_ffa.md) | yinhai_ffa | FFA | 812 | 34 | 52 | 0 |
|
| 120 |
| [YinhaiUBM](cards/YinhaiUBM.md) | YinhaiUBM | UBM | 75,720 | 4,012 | 3,857 | 0 |
|
| 121 |
-
| **Total** | | |
|
| 122 |
|
| 123 |
Split rules: `official` = official split; `official_val_from_train` = 10% of the official training set (by patient) used as
|
| 124 |
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
|
|
|
|
| 12 |
path: data/*/validation-*.parquet
|
| 13 |
- split: test
|
| 14 |
path: data/*/test-*.parquet
|
| 15 |
+
- config_name: BUltrasound
|
| 16 |
+
data_files:
|
| 17 |
+
- split: train
|
| 18 |
+
path: data/BUltrasound/train-*.parquet
|
| 19 |
+
- split: validation
|
| 20 |
+
path: data/BUltrasound/validation-*.parquet
|
| 21 |
+
- split: test
|
| 22 |
+
path: data/BUltrasound/test-*.parquet
|
| 23 |
- config_name: chengdu_ffa
|
| 24 |
data_files:
|
| 25 |
- split: train
|
|
|
|
| 96 |
|
| 97 |
# OphGen — curated institutional ophthalmic datasets
|
| 98 |
|
| 99 |
+
Curated copies of 10 institutional ophthalmic imaging datasets, prepared for the OphGen ophthalmic generative-model project:
|
| 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 |
|
|
|
|
| 106 |
|
| 107 |
```python
|
| 108 |
from datasets import load_dataset
|
| 109 |
+
ds = load_dataset("Kaphathy/OphGen-Private", "BUltrasound") # one dataset, its own splits and labels
|
| 110 |
+
side = load_dataset("Kaphathy/OphGen-Private", "BUltrasound-sidecar") # its other annotation files (path, size, bytes), if any
|
| 111 |
pre = load_dataset("Kaphathy/OphGen-Private", "all") # every dataset (default config), e.g. for pretraining
|
| 112 |
```
|
| 113 |
|
|
|
|
| 117 |
|
| 118 |
| Config | Dataset | Modalities | Train | Validation | Test | Tasks |
|
| 119 |
| :--- | :--- | :--- | ---: | ---: | ---: | ---: |
|
| 120 |
+
| [BUltrasound](cards/BUltrasound.md) | BUltrasound | Conventional ocular B-scan ultrasonography | 101,473 | 5,689 | 5,574 | 0 |
|
| 121 |
| [chengdu_ffa](cards/chengdu_ffa.md) | chengdu_ffa | FFA | 3,519 | 473 | 934 | 1 |
|
| 122 |
| [eyFRDCls](cards/eyFRDCls.md) | eyFRDCls | CFP | 418 | 53 | 139 | 1 |
|
| 123 |
| [FQ_Datasets](cards/FQ_Datasets.md) | FQ_Datasets | CFP | 2,681 | 402 | 802 | 1 |
|
|
|
|
| 127 |
| [trhd3_FRD_Cls66_Train](cards/trhd3_FRD_Cls66_Train.md) | trhd3_FRD_Cls66_Train | CFP | 7,780 | 1,106 | 2,176 | 1 |
|
| 128 |
| [yinhai_ffa](cards/yinhai_ffa.md) | yinhai_ffa | FFA | 812 | 34 | 52 | 0 |
|
| 129 |
| [YinhaiUBM](cards/YinhaiUBM.md) | YinhaiUBM | UBM | 75,720 | 4,012 | 3,857 | 0 |
|
| 130 |
+
| **Total** | | | 243,440 | 15,214 | 17,970 | |
|
| 131 |
|
| 132 |
Split rules: `official` = official split; `official_val_from_train` = 10% of the official training set (by patient) used as
|
| 133 |
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
|