cSeg2022 is a dataset focused on the segmentation of T1-weighted MRI images of 24-month-old infants, specifically targeting the detailed segmentation of the cerebellar structure. Compared to other brain segmentation datasets, cSeg2022 is quite similar to iSeg, which was initiated by the University of North Carolina at Chapel Hill in 2017 (both involve complete segmentation of brain structures and both categorize into cerebrospinal fluid, grey matter, and white matter). The important difference is that this dataset focuses only on cerebellar structures, and the chosen subjects are 24-month-old infants. Clinically, this task does not have significant differences; overall, it can be used as a supplemental dataset to iSeg or in specific cerebellar research.
In medical image segmentation, the study of brain structures is particularly crucial, with the significance of infant brain structure research being self-evident. The cerebellum is a key brain structure that develops rapidly during the early postnatal stages of infants, and its accurate segmentation into white matter, grey matter, and cerebrospinal fluid is vital for revealing the early development process of the cerebellum. However, compared to the adult cerebellum, segmentation studies for the infant cerebellum are noticeably lacking. Given the extremely low tissue contrast and severe partial volume effects in infant cerebellum MRI, both manual and automatic segmentation face significant challenges.
| Dimensions | Modality | Task Type | Anatomical Structures | Anatomical Area | Number of Categories | Data Volume | File Format |
|---|---|---|---|---|---|---|---|
| 3D | MR | Segmentation | Cerebrospinal fluid, grey matter, white matter. | Brain | 3 | 13 | .nii.gz |
You can also write down the resolution details such as size (x, y, z) and spacing information about the dataset.
For example:
| Dataset Statistics | spacing (mm) | size |
|---|---|---|
| min | (0.80, 0.80, 0.80) | (300, 300, 300) |
| median | (0.80, 0.80, 0.80) | (300, 300, 300) |
| max | (0.80, 0.80, 0.80) | (300, 300, 300) |
| Anatomical Structure | Cerebrospinal Fluid | Gray Matter | White Matter |
|---|---|---|---|
| Cases | 13 | 13 | 13 |
| Coverage | 100% | 100% | 100% |
| Mean Volume (cm³) | 22 | 50 | 31 |
| Median Volume (cm³) | 27 | 67 | 36 |
| Max Volume (cm³) | 34 | 81 | 49 |
Full structure visualization.
Cerebrospinal fluid visualization.
Grey matter visualization.
White matter visualization.
The dataset file structure is as follows, following the nnU-Net style of data organization, containing 'imagesTr' and 'labelsTr' folders for storing images and labels, respectively:
cSeg-2022
├── imagesTr
│ ├── subject-1.nii.gz
│ ├── ...
│ ├── subject-13.nii.gz
├── labelsTr
│ ├── subject-1.nii.gz
│ ├── ...
│ ├── subject-13.nii.gz
├── dataset.json
Yue Sun (University of North Carolina at Chapel Hill)
Limei Wang (University of North Carolina at Chapel Hill)
Dr. Li Wang (University of North Carolina at Chapel Hill)
Dr. Gang Li (University of North Carolina at Chapel Hill)
Dr. Valerie Jewells (University of North Carolina at Chapel Hill)
Dr. Ian H. Gotlib (Stanford University)
Dr. Kathryn Leigh Humphreys (Vanderbilt University)
Dr. Weili Lin (University of North Carolina at Chapel Hill)
Official Website: https://tarheels.live/cseg2022/
Download Link: https://tarheels.live/cseg2022/data/
Article Address: TBD
Publication Date: April, 2022.
TBD
Original introduction article is here.