Shared-boundary fusion for estimation of noisy multi-modality atherosclerotic plaque imagery

Our work focuses on applying boundary-preserving smoothing techniques to the fusion of multiple image modalities in an effort to improve image-based classification of atherosclerotic lesions. No single imaging modality has yet demonstrated the ability to reliably detect "vulnerable" lesions. We present an approach for estimating multi-modality biomedical imagery when tissue classes are sharply delimited by boundaries. Our approach is based on the Mumford-Shah framework for edge-preserving smoothing. We exploit this framework to fuse heterogeneous sensing modalities that image unrelated physicochemical parameters of a piecewise-homogeneous tissue field. We demonstrate this approach by fusing boundary field estimates from MR and CT atherosclerotic lesion imagery into a single estimated underlying tissue boundary field, while simultaneously estimating the original imagery to better estimate tissue characteristics and structure.