scan.gif: CT slice through the head of a metacarpal bone.



This page contains several datasets of human bony structures.
These images have been made public because of the difficulties young researchers may sometimes have to get access to data of a good quality to perform their research.

Scientific purpose: There is absolutly no copyright on the images when they are used for scientific purpose. The authors only ask you to mention the source of the images when the results are published (paper, poster, oral communication). See beside for details.

Commercial purpose: These data were collected using part of the research funding of the authors. So prior to any commercial use, and as gentlemen, you are kindly asked to contact the authors (see below) to find an agreement.

Please, quote the following in the acknowledgement of any publication, poster or oral communication that used the data available on this page. Thank you!!

"The image datasets used in this experiment were from the Laboratory of Human Anatomy and Embryology, University of Brussels (ULB), Belgium".

For any further information, please contact:
Serge VAN SINT JAN, PhD. Laboratory of Human Anatomy and Embryology (CP 619). Faculty of Medicine. University of Brussels (ULB). Route de Lennik 808. 1070 Brussels. Belgium. Fax: +32-2-555 63 78



At the time 3D-reconstruction and virtual environment techniques are booming, young (and older!) scientists often have difficulties to get good quality pictures because, well, we are living in a world where time is expensive and it is sometimes hard to convince somebody to spend part of his precious time to collect data for a particular experiment. Furthermore, imaging installations are almost always used for clinical investigations (which it's normal!).
Some years ago, we were lucky enough to meet Dr. M. Lemort (Department of Radiology, Bordet Institute, University of Brussels (ULB), Belgium) who welcomed us with opened arms and allowed us to use the imaging facilities of his department (in the evening after the clinical program, of course, but in a trust relationship that must be underlined!). Consequently, we'd like to express here all our gratitude to Dr. Lemort who made possible to put good (we hope!) data freely available on the Net for scientific purposes.
We would also greatly appreciate to receive some feedback from your work. Please, send any abstract, offprint, or simply comment or note at the above-mentioned address ( Thank you!!

The Laboratory of Human Anatomy and Embryology, Faculty of Medicine, University of Brussels (ULB), Belgium

Image and CT-Scan specifications

The images are in DICOM format which is the standard for medical imaging.

The files have been compressed using WinZip (available at the WinZip website, click here). For unzipping on UNIX computers, try "unzip filename". If unzip is not installed, here is an SGI executable. For other systems, try this link to C source code .

DICOM file reader and code are freely available at several locations.
A good reader is DICOM2 (click here). It allows converting DICOM files to several formats and listing their file header.

A huge library of routines for DICOM file manipulation can be found by clicking here.

The Ct-Scan installation used to collect the data was a Helicoidal Twin from Elscint (Haifa, Israel).

Image parameters

The pages with the image file link (see The images below), also shows several parameters about, e.g., image dimensions, acquisition parameters, and so on. The table below gives our definition of those parameters (you could possibly meet the same parameters with other names elsewhere). Some are useful, others not and are given for information only.

Sequence For some long bones, the data have been acquired in several shots. Generally, the images with the epiphysis (i.e., part of the bone with a joint surface) show an higher resolution and a lower slice thickness than the diaphysis (i.e., central part of the bone).
So, for example, the full humeral bone is available through 3 datasets which sequence have been indexed 1, 2, 3 respectively.
!!The 3 shots (3 in that particular case) have been obtained independently, but automatically by the CT installation. So, the "imaged" bone was not moved between the sequences and you can use the 3 datasets to build one full humeral bone.
Of course you need to take care of the difference in the slice increment between the dateset. Also, the Z position of the first and last slice (see below) is given when there are several datasets for one bone.
When only 1 sequence is indicated, it means that the bone data have been obtained in one shot.
Structure Anatomical structure name.
Orientation Left or right. If no orientation is given: the structure is odd (e.g., the sacral bone).
Pitch We used a helicoidal sequences. So, each sequence is characterized by a pitch.
Thickness (mm) Slice thickness.
Increment (mm) Slice spacing.
First Z (mm) Z position of the first slice of the current set. This is useful when one dataset is made of several sequences.
Last Z (mm) Z position of the last slice of the current set. This is useful when one dataset is made of several sequences.
Nb of slices No comment.
Matrix (pixel) Rectangular dimension of the images (usually 512).
Dimension (mm) Real image dimension (usually 180 or 250).
Voltage (KV) No comment.
Amperage (mA) No comment.
Duration (sec) Image acquisition time.
Size (MB) Size of the uncompressed dataset.
Download (...Kb) Link to the image files (size in Kb of the compressed file).

The images

isblogo.small.gif Back to ISB Biomechanical Data Sources

Last Update: Aug 15 1998 by S. Van Sint Jan.