Repository for Image Data

Harris MD, Anderson AE, Henak CR, Ellis BJ, Peters CL, Weiss JA:  Finite element prediction of cartilage contact stresses in normal human hips.  Journal of Orthopaedic Research, 30(7),  2012.


It is difficult to obtain high resolution image data of human diarthrodial joints.  This is due to the effort required to obtain IRB approval, to recruit subjects and to perform the image acquisition.    Therefore, we have made CT arthrogram image datasets available from the 10 subjects included in our finite element analysis of cartilage contact stresses.  We hope that the availability of these image datasets will assist other investigators who wish to study the articular cartilage and/or mechanics of the normal human hip joint.

Accessing the Image Data

Separate files are available for each of the 10 subjects.  Each file is a Linux tape archive (tar) file, compressed with GNU Zip (gzip).  Thus each file has the ending “.tar.gz”.  The size of each file is approximately 3 GB.
Under Linux or other Unix operating systems, each individual download can be uncompressed and extracted using the commands:

  • gunzip .tar.gz
  • tar – xvf .tar

This will create a subdirectory containing all of the image data files for the particular subject in DICOM format, as well as a copy of the license agreement and this file.  Each image file represents one slice of the image dataset.
Both gzip and tar are available for Windows, and there are a number of free software tools that can extract .tar.gz files under Windows.
The image files have the extension “.dcm”.  Many image processing, segmentation and visualization programs can open these DICOM files directly.  The free software ImageJ is one option:

To open an image stack in Image J, choose “Import” from the “File” menu, and then choose “Image Sequence”.  Choose the first image in the stack and then click “Open”.

Subject Information

Volunteers were recruited to match the age, weight, and body mass index of patients with acetabular dysplasia commonly treated in our clinic.  The basic characteristics of these volunteers are listed in the table below.

Subject #AgeSexWeight (kg)Height (cm)BMI

Image Acquisition

CT images were acquired at the University of Utah Hospital.  One hip from each subject was selected randomly to receive a CT arthrogram.  The hip capsule was injected with ~20 ml of a diluted contrast agent (2:1 lidocaine to OMNIPAQUE® 350, GE Healthcare Inc, Princeton, NJ) under fluoroscopic guidance.  Multi-detector CT scans of the entire pelvis and both femurs were obtained within 10 minutes of injection (120 kVp, 100-400 mAs, 512×512 matrix, 1.0 pitch, 300-400 mm FOV, 0.75-1.0 mm slice thickness).  Joint traction was applied during the scan using a hare traction device to ensure that the contrast agent filled the joint space.  The images below indicate the bone and cartilage regions in the image data.

Resampled Image Information

CT images were resampled using a Lanczos filter kernel to improve the resolution of the segmentation masks.  These resampled images are provided on the web page.

Subject #Resampled in-plane resolutionOriginal
in-plane resolution
in-plane pixel spacing (mm)
Resampled slice thickness (mm)Number of slicesOriginal slice thickness (mm)Bit depth
JOR011536 x 1536512×5120.21875 x 0.218750.233310380.7016
JOR021536 x 1536512×5120.22526 x 0.225260.233310650.7016
JOR031536 x 1536512×5120.23307 x 0.233070.233310720.7016
JOR041536 x 1536512×5120.23372 x 0.233720.33338051.016
JOR051536 x 1536512×5120.21875 x 0.218750.33337691.016
JOR061536 x 1536512×5120.23697 x 0.236970.33337791.016
JOR071536 x 1536512×5120.21549 x 0.215490.33338261.016
JOR081536 x 1536512×5120.21744 x 0.217440.33337561.016
JOR091536 x 1536512×5120.26367 x 0.263670.33337351.016
JOR101536 x 1536512×5120.24674 x 0.246740.33337231.016