Fetal Tissue Annotation and Segmentation Dataset FeTA
The Fetal Tissue Annotation and Segmentation Dataset (FeTA) is a unique collection of super-resolution reconstructed fetal cerebral MRI scans and ground truth manual annotations of brain tissues. The dataset was created to foster machine learning algorithm development and testing. We provide fetal cerebral MRI and corresponding annotation of seven tissue types. The dataset serves the purpose of benchmarking fetal cerebral MRI segmentation methods, and therefore will be used in conjunction with upcoming challenges. The dataset is split into training (40 subjects) and an independent testing set (10 subjects). For the test set, we will not provide the ground truth labels, this way, the accuracy for new observations can be estimated in an unbiased way and compared across groups who wish to participate in the segmentation challenges. The dataset structure is compatible with BIDS 1.4.
Specifications
License: