A Meta Registration Framework for Lesion Matching ⋆ Sharmishtaa Seshamani 1 , Purnima Rajan 1 , Rajesh Kumar 1 , Hani Girgis 1 , Themos Dassopoulos 3 , Gerard Mullin 2 , and Gregory Hager 1 1 Department of Computer Science, Johns Hopkins University, Baltimore, MD sharmi@jhu.edu 2 Johns Hopkins Hospital, Baltimore, MD 3 Washington University, School of Medicine St. Louis, MO Abstract. A variety of pixel and feature based methods have been proposed for registering multiple views of anatomy visible in studies obtained using diagnostic, minimally invasive imaging. A given regis- tration method may outperform another depending on anatomical vari- ations, imaging conditions, and imaging sensor performance, and it is often difficult a priori to determine the best registration method for a particular application. To address this problem, we propose a registra- tion framework that pools the results of multiple registration methods using a decision function for validating registrations. We refer to this as meta registration. We demonstrate that our framework outperforms several individual registration methods on the task of registering multi- ple views of Crohn’s disease lesions sampled from a Capsule Endoscopy (CE) study database. We also report on preliminary work on assessing the quality of registrations obtained, and the possibility of using such assessment in the registration framework. 1 Introduction Minimally invasive diagnostic imaging methods such as flexible endoscopy, and wireless capsule endoscopy (CE) often present multiple views of the same anatomy. Duplication issues are particularly severe in the case of CE, where peristaltic propulsion may lead to duplicate information over several consecutives frames, and also several minutes apart. This may be difficult to detect, since each individual image captures only a small portion of anatomical surface due to lim- ited working distance of these devices, providing relatively little spatial context. Given the relatively large anatomical surfaces (e.g. the Gastrointestinal tract (GI)) to be inspected, it is important to identify duplicate information as well as to present all available views of anatomical and disease views to the clinician for improving consistency, efficiency and accuracy of diagnosis and assessment. ⋆ Supported in part by National Institutes of Health with Grant 5R21EB008227-02 and Johns Hopkins University internal funds. G.-Z. Yang et al. (Eds.): MICCAI 2009, Part I, LNCS 5761, pp. 582–589, 2009. c  Springer-Verlag Berlin Heidelberg 2009