Patient-Specific Manifold Embedding of Multispectral Images Using Kernel Combinations
This paper presents a framework that optimizes kernel-based manifold embedding for the characterization of multispectral image data. The hypothesis is that data manifolds corresponding to high-dimensional images can have varying characteristics and types of nonlinearity. As a result, kernel functions must be selected from a wide range of transformations and tuned on an image- and patient-basis. To this end, we introduce a new measure to assess the quality of the kernel transformations that takes into account both local and global relationships in nonlinear manifolds. Furthermore, the calculated measures for each kernel are used to combine the different kernel transformations further highlight the tissue constituents in all regions of the image. Validation with phantom and real multispectral image data shows improvement in the visualization and characterization of the tissue constituents.
KeywordsMultispectral Image Class Separation Kernel Principal Component Analysis Tissue Constituent Kernel Transformation
Unable to display preview. Download preview PDF.
- 3.Hordley, S., Finalyson, G., Morovic, P.: A Multispectral Image Database and its Application to Image Rendering Across Illumination. In: Proc. 3rd Int. Conf. on Image and Graphics (ICIG), Hong-Kong, China, pp. 394–397 (2004), http://www2.cmp.uea.ac.uk/Research/compvis/MultiSpectralDB.htm
- 6.Maaten, L.V.D., Postma, E., Herik, H.V.D.: Dimensionality reduction: a comparative review. Tech. Rep. TiCC-TR 2009-05, Tilburg University, Tilburg, The Netherlands (2009)Google Scholar
- 7.Menze, B., Jakab, A., Bauer, S., Reyes, M., Prastawa, M., van Leemput, K.: Challenge on Multimodal Brain Tumor Segmentation (MICCAI-BRATS). In: Proc. MICCAI 2012 (2012), http://www2.imm.dtu.dk/projects/BRATS2012
- 8.Tanabe, H., Ho, T.B., Nguyen, C.H., Kawasaki, S.: Simple but effective methods for combining kernels in computational biology. In: IEEE Int. Conf. on Research, Innovation and Vision for the Future (RIVF), pp. 71–78 (2008)Google Scholar