Multimodal Registration Using Patch Algorithm
Images of different modality are registered using a unimodal image registration program (“Patch Algorithm”). An affine model for the displacement field is considered. Prior to registration the images are convolved with a set of Gabor filter (quadrature) pairs tuned to certain orientations and scales, and a gradient of local phase (local frequency) is computed for the filtered images. Local frequency representation of an image is relatively insensitive to changes in illumination conditions, and may enable common features between images of different modality to be captured. To demonstrate the algorithm, a series of 2- dimensional T2-weighted magnetic resonance images of brain was registered against a T1-weighted reference image. Also, registration of a near infrared hyperspectral datacube was shown to improve recovery of localized skin chromophore features for facial imaging, demonstrating its potential to obtain high spatial resolution diagnostic information.
KeywordsMutual Information Image Registration Gabor Filter Local Phase Acne Lesion
Unable to display preview. Download preview PDF.
- 6.Zhilkin, P., Alexander, M.E.: Nonlinear Registration of 3D Images Using Patch Algorithm. In: ISMRM, Honolulu, Hawaii, USA, May 18-24 (2002)Google Scholar
- 10.Cocosco, C.A., Kollokian, V., Kwan, R.K.-S., Evans, A.C.: BrainWeb: Online Interface to a 3D MRI Simulated Brain Database. NeuroImage 5(4), part 2/4, S425, (1977); Proceedings of 3-rd International Conference on Functional Mapping of the Human Brain, Copenhagen (May 1997)Google Scholar
- 12.Kollias, N., Stamatas, G.N.: Optical non-invasive approaches to diagnosis of skin diseases. Journal of investigative dermatology, The Symposium proceedings the Society for Investigative Dermatology, Inc. and European Society for Dermatological Research 7(1), 64–75 (2002)Google Scholar
- 13.Mansfield, C.D., Alexander, M.E.: Registration of Thermal and Near Infrared Hyperspectral Datacubes for Recovery of Localized Thermal or Hemodynamic Features. manuscript submitted to IEEE Transactions on Medical Imaging (2003)Google Scholar
- 14.Alexander, M.E., Summers, A.R.: Fast Registration Algorithm Using a Variational Principle for Mutual Information. In: Proceedings of SPIE Medical Imaging Conference, San Diego, February 15-20 (2003) (to appear)Google Scholar