Scan integration as a labelling problem

• We propose a novel method for multi-view scan integration where we solve it as a labelling problem.
• Our labelling-based method is essentially a selection strategy.
• We formulate the labelling via a higher-order Markov Random Field.
• The proposed method is more robust to registration errors than previous methods, producing significantly improved integration.

Integration is a crucial step in the reconstruction of complete 3D surface model from multiple scans. Ever-present registration errors and scanning noise make integration a nontrivial problem. In this paper, we propose a novel method for multi-view scan integration where we solve it as a labelling problem. Unlike previous methods, which have been based on various merging schemes, our labelling-based method is essentially a selection strategy. The overall surface model is composed of surface patches from selected input scans. We formulate the labelling via a higher-order Markov Random Field (MRF) which assigns a label representing an index of some input scan to every point in a base surface. Using a higher-order MRF allows us to more effectively capture spatial relations between 3D points. We employ belief propagation to infer this labelling and experimentally demonstrate that this integration approach provides significantly improved integration via both qualitative and quantitative comparisons.