Update README.md
Browse files
README.md
CHANGED
@@ -1,31 +1,39 @@
|
|
1 |
---
|
2 |
license: cc-by-4.0
|
3 |
---
|
|
|
4 |
# ReefNet Dataset
|
5 |
|
6 |
-
The ReefNet dataset is a comprehensive collection of annotated coral images
|
|
|
|
|
7 |
|
8 |
## Dataset Structure
|
9 |
|
10 |
The dataset is organized as follows:
|
11 |
|
12 |
-
- `
|
13 |
-
- Contains the annotations
|
|
|
|
|
|
|
14 |
|
15 |
-
- `
|
16 |
-
- Contains the annotations
|
|
|
17 |
|
18 |
-
- `
|
19 |
-
-
|
|
|
20 |
|
21 |
-
- `
|
22 |
-
-
|
23 |
|
24 |
-
- `
|
25 |
-
- Provides an overview of the dataset
|
26 |
|
27 |
-
- `
|
28 |
-
-
|
29 |
|
30 |
## Download Instructions
|
31 |
|
|
|
1 |
---
|
2 |
license: cc-by-4.0
|
3 |
---
|
4 |
+
---
|
5 |
# ReefNet Dataset
|
6 |
|
7 |
+
The ReefNet dataset is a comprehensive, curated collection of (annotated) coral reef images sourced from the CoralNet platform.
|
8 |
+
The ReefNet dataset is designed for training and evaluating machine learning models for coral classification on a global scale.
|
9 |
+
Furthermore, ReefNet's consolidated labelset and rich metadata facilitate straigforward extraction of custom datasets to support coral classification tasks across different geographic regions.
|
10 |
|
11 |
## Dataset Structure
|
12 |
|
13 |
The dataset is organized as follows:
|
14 |
|
15 |
+
- `All_ReefNet_annotations.csv` (2.98 GB)
|
16 |
+
- Contains the all annotations available in the main ReefNet dataset, together with the associated metadata.
|
17 |
+
|
18 |
+
- `Metadata_All_ReefNet_annotations.xlsx` (11 KB)
|
19 |
+
- Contains metadata for the `All_ReefNet_annotations.csv` file, describing the meaning of the column names.
|
20 |
|
21 |
+
- `ReefNet_RSG-TestDataset_annotations.csv` (2.42 MB)
|
22 |
+
- Contains the annotations of the Red Sea Global (RSG) test dataset. This newly released dataset is used in experiment 2 of the main paper as the test dataset.
|
23 |
+
This dataset was collected by Red Sea Global (RSG) in the Al-Wajh lagoon in the Red Sea (26.4°N, 36.4°E) in 2022 and was previously stored as a private source within CoralNet.
|
24 |
|
25 |
+
- `ReefNet_RSG-TestDataset_images.zip` (1.21 GB)
|
26 |
+
- Contains the annotations of the Red Sea Global (RSG) test dataset. This newly released dataset is used in experiment 2 of the main paper as the test dataset.
|
27 |
+
This dataset was collected by Red Sea Global (RSG) in the Al-Wajh lagoon in the Red Sea (26.4°N, 36.4°E) in 2022 and was previously stored as a private source within CoralNet.
|
28 |
|
29 |
+
- `Metadata_ReefNet_by_CoralNet_sources.xlsx` (33 KB)
|
30 |
+
- Provides an overview of the CoralNet sources the ReefNet dataset was compiled from, including attribution to the original creators and metadata such as the Geographic location or average image resolution.
|
31 |
|
32 |
+
- `ReefNet_labelmapping.xlsx` (79 KB)
|
33 |
+
- Provides an overview of the labelset mapping that was performed to consolidate the ReefNet dataset from the separate CoralNet sources.
|
34 |
|
35 |
+
- `ReefNet_RSG-TestDataset_labelmapping.xlsx` (33 KB)
|
36 |
+
- Provides an overview of the labelset mapping that was performed for experiment 2 in the main paper to merge labels from the main ReefNet dataset to the RSG-TestDataset.
|
37 |
|
38 |
## Download Instructions
|
39 |
|