Dataset¶

In the challenge, the participants will be provided the US nodular thyroid dataset supplied by Shanghai Ruijin Hospital. This data comes from the Chinese Medical Ultrasound Artificial Intelligence Alliance (CMUAA) initiated by doctor JianQiao Zhou. The dataset includes training set images and testing set images in grayscale and .png extension from different equipments and sizes. All personal labels of scan images were removed in order to protect patients' privacy.
    The training set: 3644 US nodular thyroid images of 3644 patients. The annotation of nodules is labeled by experienced doctors. The annotation images for segmentation task are binary images in which pixels are either 255 for the foreground or 0 for the background. The annotation images named as “xxx.png”. Where “xxx” presents patient ID (from 001 to 3644). The annotation of the classification task is in CSV file with a header of ID (denotes for patient ID) and CATE (0 refers to Benign, and 1 refers to Malignant).

An example of a training US image of thyroid nodule with the corresponding mask

According to the MICCAI proposal (March 19,2020), Participation policies, target="_blank">Point f: 

The publish right of this DATASET is limited to the purpose of this CHALLENGE ONLY due to the ethical approval was obtained. NO publication rights will be given on the dataset images apart from this challenge. The participants will NOT be able to use the dataset images in any other publication or study.

We apologize for any inconvience this might cause.