updated README.md
Browse files
README.md
CHANGED
|
@@ -28,23 +28,42 @@ Original data are available on [Zenodo](https://zenodo.org/records/8009680). Mor
|
|
| 28 |
<figcaption>Example MRI scan with segmentation masks</figcaption>
|
| 29 |
</figure>
|
| 30 |
|
| 31 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
|
| 33 |
1. [Getting Started](https://huggingface.co/datasets/cdoswald/SPIDER#getting-started)
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
3. [
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
|
| 49 |
First, you will need to install the following dependencies:
|
| 50 |
|
|
@@ -62,13 +81,7 @@ dataset = load_dataset("cdoswald/SPIDER, name="default", trust_remote_code=True)
|
|
| 62 |
More detailed examples for [loading](tutorials/load_data.ipynb) the dataset with different configurations
|
| 63 |
and using the dataset for [segmentation tasks](tutorials/segment_anything.ipynb) are provided in the [tutorials](tutorials) folder.
|
| 64 |
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
- **Paper:** [Lumbar spine segmentation in MR images: a dataset and a public benchmark](https://www.nature.com/articles/s41597-024-03090-w)
|
| 68 |
-
- **Repository:** [Zenodo](https://zenodo.org/records/8009680)
|
| 69 |
-
- **Grand Challenge:** [SPIDER Grand Challenge](https://spider.grand-challenge.org/)
|
| 70 |
-
|
| 71 |
-
### Dataset Summary
|
| 72 |
|
| 73 |
The dataset includes 447 sagittal T1 and T2 MRI series collected from 218 patients across four hospitals.
|
| 74 |
Segmentation masks indicating the vertebrae, intervertebral discs (IVDs), and spinal canal are also included.
|
|
@@ -78,7 +91,7 @@ In addition to MR images and segmentation masks, additional metadata (e.g., scan
|
|
| 78 |
patient characteristics (biological sex and age, when available), and radiological gradings indicating specific degenerative
|
| 79 |
changes can be loaded with the corresponding image data.
|
| 80 |
|
| 81 |
-
|
| 82 |
|
| 83 |
This version of the SPIDER dataset (i.e., available through the HuggingFace `datasets` library) differs from the original
|
| 84 |
data available on [Zenodo](https://zenodo.org/records/8009680) in two key ways:
|
|
@@ -98,8 +111,7 @@ and 39 (15%) studies for the SPIDER Grand Challenge [hidden test set](https://sp
|
|
| 98 |
their models to the SPIDER Grand Challenge, the original 218 studies that comprised the public training/validation set were further partitioned using a 60%/20%/20% split. The original split
|
| 99 |
for each study (i.e., training or validation set) is recorded in the `OrigSubset` variable in the study's linked metadata.
|
| 100 |
|
| 101 |
-
|
| 102 |
-
## Dataset Structure
|
| 103 |
|
| 104 |
### Data Instances
|
| 105 |
|
|
@@ -192,7 +204,7 @@ An additional hidden test set provided by the paper authors
|
|
| 192 |
(i.e., not available via HuggingFace) is available on the
|
| 193 |
[SPIDER Grand Challenge](https://spider.grand-challenge.org/spiders-challenge/).
|
| 194 |
|
| 195 |
-
|
| 196 |
|
| 197 |
> Standard sagittal T1 and T2 image resolution ranges from 3.3 x 0.33 x 0.33 mm to 4.8 x 0.90 x 0.90 mm.
|
| 198 |
> Sagittal T2 SPACE sequence images had a near isotropic spatial resolution with a voxel size of 0.90 x 0.47 x 0.47 mm.
|
|
@@ -203,7 +215,7 @@ for compatibility with the HuggingFace `datasets` library. If you want to use th
|
|
| 203 |
load the original images from the local cache indicated in each example's `image_path` and `mask_path` features.
|
| 204 |
See the data loading [tutorial](tutorials/load_data.ipynb) for more information.
|
| 205 |
|
| 206 |
-
|
| 207 |
|
| 208 |
### License
|
| 209 |
|
|
|
|
| 28 |
<figcaption>Example MRI scan with segmentation masks</figcaption>
|
| 29 |
</figure>
|
| 30 |
|
| 31 |
+
# Dataset Description
|
| 32 |
+
|
| 33 |
+
- **Paper:** [Lumbar spine segmentation in MR images: a dataset and a public benchmark](https://www.nature.com/articles/s41597-024-03090-w)
|
| 34 |
+
- **Repository:** [Zenodo](https://zenodo.org/records/8009680)
|
| 35 |
+
- **Grand Challenge:** [SPIDER Grand Challenge](https://spider.grand-challenge.org/)
|
| 36 |
+
|
| 37 |
+
# Table of Contents (TOC)
|
| 38 |
|
| 39 |
1. [Getting Started](https://huggingface.co/datasets/cdoswald/SPIDER#getting-started)
|
| 40 |
+
|
| 41 |
+
2. [Dataset Summary](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 42 |
+
|
| 43 |
+
3. [Data Modifications](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 44 |
+
|
| 45 |
+
4. [Dataset Structure](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 46 |
+
|
| 47 |
+
- [Data Instances](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 48 |
+
|
| 49 |
+
- [Data Format/Fields](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 50 |
+
|
| 51 |
+
- [Data Splits](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 52 |
+
|
| 53 |
+
5. [Image Resolution](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 54 |
+
|
| 55 |
+
6. [Additional Information](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 56 |
+
|
| 57 |
+
- [License](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 58 |
+
|
| 59 |
+
- [Citation](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 60 |
+
|
| 61 |
+
- [Disclaimer](https://huggingface.co/datasets/cdoswald/SPIDER#dataset-summary)
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
<br>
|
| 65 |
+
|
| 66 |
+
# Getting Started
|
| 67 |
|
| 68 |
First, you will need to install the following dependencies:
|
| 69 |
|
|
|
|
| 81 |
More detailed examples for [loading](tutorials/load_data.ipynb) the dataset with different configurations
|
| 82 |
and using the dataset for [segmentation tasks](tutorials/segment_anything.ipynb) are provided in the [tutorials](tutorials) folder.
|
| 83 |
|
| 84 |
+
# Dataset Summary
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 85 |
|
| 86 |
The dataset includes 447 sagittal T1 and T2 MRI series collected from 218 patients across four hospitals.
|
| 87 |
Segmentation masks indicating the vertebrae, intervertebral discs (IVDs), and spinal canal are also included.
|
|
|
|
| 91 |
patient characteristics (biological sex and age, when available), and radiological gradings indicating specific degenerative
|
| 92 |
changes can be loaded with the corresponding image data.
|
| 93 |
|
| 94 |
+
# Data Modifications
|
| 95 |
|
| 96 |
This version of the SPIDER dataset (i.e., available through the HuggingFace `datasets` library) differs from the original
|
| 97 |
data available on [Zenodo](https://zenodo.org/records/8009680) in two key ways:
|
|
|
|
| 111 |
their models to the SPIDER Grand Challenge, the original 218 studies that comprised the public training/validation set were further partitioned using a 60%/20%/20% split. The original split
|
| 112 |
for each study (i.e., training or validation set) is recorded in the `OrigSubset` variable in the study's linked metadata.
|
| 113 |
|
| 114 |
+
# Dataset Structure
|
|
|
|
| 115 |
|
| 116 |
### Data Instances
|
| 117 |
|
|
|
|
| 204 |
(i.e., not available via HuggingFace) is available on the
|
| 205 |
[SPIDER Grand Challenge](https://spider.grand-challenge.org/spiders-challenge/).
|
| 206 |
|
| 207 |
+
# Image Resolution
|
| 208 |
|
| 209 |
> Standard sagittal T1 and T2 image resolution ranges from 3.3 x 0.33 x 0.33 mm to 4.8 x 0.90 x 0.90 mm.
|
| 210 |
> Sagittal T2 SPACE sequence images had a near isotropic spatial resolution with a voxel size of 0.90 x 0.47 x 0.47 mm.
|
|
|
|
| 215 |
load the original images from the local cache indicated in each example's `image_path` and `mask_path` features.
|
| 216 |
See the data loading [tutorial](tutorials/load_data.ipynb) for more information.
|
| 217 |
|
| 218 |
+
# Additional Information
|
| 219 |
|
| 220 |
### License
|
| 221 |
|