Skip to main content

Example Datasets

Below we have a selection of available example datasets for you to download. We will help you go through the steps for downloading them so that they are ready to be added as image sets to projects for labelling.

3D DICOM dataset​

Non-Small Cell Lung Cancer CT Classification

Here are the following steps to download this dataset:

  1. Download the images manifest file from the Cancer Imaging Archive here [1]: NSCLC-Radiomics-Interobserver1. This will be a small file with the extension .tcia. There should be 3 DICOM series in this dataset.

  2. Install the NBIA Data Retriever. Instructions can be found here.

  3. Open the .tcia file with the NBIA Data Retriever application, then follow the steps to download the DICOM dataset.

  4. Move the folder NSCLC-Radiomics-Interobserver1, which is inside the downloaded folder manifest-####, to your user's home directory (e.g. /home/user1/).

2D DICOM dataset​

Lung X-Ray Classification

Here are the steps to download this dataset:

  1. Download the data folder here: download dataset.

  1. Extract the downloaded compressed file pneu_2d_dcm_x15-####.zip.

  2. The extracted folder contains the directory pneu_2d_dcm_x15. Move this to your user's home directory (e.g. /home/user1/).

  3. The images contained in pneu_2d_dcm_x15 can be used as an example dataset to be labelled:

  • The folder images, which contains 15 chest x-ray images.
  • The ground truth csv file: ground_truth_rel.csv.
  1. We have attached the ground truth CSV, with the relative paths to the individual images for curiosity purposes but these are not needed for the labelling product.

Once at least one of these datasets has been downloaded, you can upload them as an image set to the labelling platform as demonstrated in this step by step guide.

Footnotes​

  1. Wee, L., Aerts, H. J.L., Kalendralis, P., & Dekker, A. (2019). Data from NSCLC-Radiomics-Interobserver1 [Data set]. The Cancer Imaging Archive. https://doi.org/10.7937/tcia.2019.cwvlpd26. ↩

  2. Shih G, et al. Augmenting the National Institutes of Health Chest Radiograph Dataset with Expert Annotations of Possible Pneumonia. https://pubs.rsna.org/doi/10.1148/ryai.2019180041 ↩