Revealing bone damage using radiographic image registration
Kauffman, Joost and Slump, Kees and Bernelot Moens, Hein (2007) Revealing bone damage using radiographic image registration. In: Third Annual IEEE BENELUX/DSP Valley Signal Processing Symposium, SPS-DARTS 2007, 21-22 March 2007, Antwerp, Belgium.
|Abstract:||Bone damage assessment is frequently applied to monitor the activity of bone degenerative diseases such as rheumatoid arthritis and osteoarthritis. For an effective treatment it is important that small changes over time can be measured. Radiographs of hands and feet are often used for such measurements. Several scoring methods exist to measure bone and joint damage , but they are subjected to inter-observer and intra-observer variability. We present a method for comparing radiographs that have been taken at different moments in time. Using the segmentation algorithm presented in earlier work  we select corresponding regions of interest surrounding the bone to be analyzed. Both image selections taken at different time-points are aligned to each other with a registration algorithm . After aligning the images, we visualize the difference by image subtraction. Since there is generally no information available about the setup of the radiographic system during both acquisitions, we compensate for differences in lighting. We do this by estimating an intensity transformation function based on the joint density function of both images. Experimental results with several follow-up radiographs show that we are able to visualize small erosions and changes in bone mineral density. To further improve the estimation of the intensity
transformation function, we plan to use a calibration object in future research.
|Item Type:||Conference or Workshop Item|
Electrical Engineering, Mathematics and Computer Science (EEMCS)
|Link to this item:||http://purl.utwente.nl/publications/67148|
|Export this item as:||BibTeX|
Daily downloads in the past month
Monthly downloads in the past 12 months
Repository Staff Only: item control page
Metis ID: 241649