Localization of fiducial skin markers in MR images using correlation pattern recognition for PET/MRI nonrigid breast image registration

• Published in: 2008 37th IEEE Applied Imagery Pattern Recognition Workshop pp. 1 - 4
• Main Authors: Walvoord, D., Baum, K.G., Helguera, M., Krol, A., Easton, R.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-10-2008
• Subjects: Biomedical imaging; Breast; Data visualization; Filters; Image registration; Magnetic resonance imaging; Pathology; Pattern recognition; Positron emission tomography; Skin
• ISBN: 1424431255; 9781424431250
• ISSN: 1550-5219; 2332-5615
• DOI: 10.1109/AIPR.2008.4906474

In most instances, multiple-modality visualization of pathologies will present advantages over single-modality studies. For many medical imaging procedures, it is desirable to produce a ldquofusedrdquo output that simultaneously exhibits characteristics of the data from each individual modality to reduce the difficulty of the decision-making process for radiologists. Preprocessing for most data fusion algorithms typically performs the necessary registration of the input data (from each modality). Fiducial markers may be used to display image locations that are common to each modality if the images exhibit very different spatial structure, as is the case with MRI and PET imagery. The process of automating the detection of these markers needs more investigation in the medical field, and the current state of the art often requires manual selection by a human observer throughout the 3-dimensional image stack. The objective of automated detection is to locate the centroid of each marker in a noisy background that contains additional objects that span a large range of intensity values. The correlation methods employed must somehow ldquonormalizerdquo the images to accommodate changes in the input image, so that ldquobrightnessrdquo of the region does not affect the correlation, thus reducing the false positive rate. The filter should accommodate within-class distortion, as the size and shape of the fiducial marker will vary through the image stack. For this work, a mean-subtracted MACH filter was constructed and applied to data that are mean-subtracted locally. The location of centroids in the output stack of correlation planes was determined by applying morphological operations to extract regions-of-interest. It is apparent that a relatively high probability of detection is obtained over a wide range of thresholds with an acceptable false positive rate.