SKM-TEA Knee MRI

The SKM-TEA dataset consists of imaging data and annotations for 155 quantitative double echo steady state MRI knee scans acquired clinically at Stanford. The data includes the raw kspace, DICOM images, segmentations of six tissues, and bounding boxes for 16 pathologies. The dataset consists of 86 scans for training, 33 scans for validation, and 36 scans for testing. All data were acquired with 2x1 parallel imaging using 8 or 15 coils with elliptical MRI sampling. Missing data was subsequently estimated using ARC (GE) parallel imaging with the GE Orchestra MATLAB SDK. This data is considered the fully-sampled kspace. DICOM images were manually segmented for articular tissue and the meniscus. The were also annotated for 16 different pathologies that were extracted and localized (3D bounding boxes) based on corresponding radiology reports. More details here: https://openreview.net/forum?id=YDMFgD_qJuA