Reassembling Thin Artifacts of Unknown Geometry

Abstract

We introduce a novel reassembly method for fragmented, thin objects that uses minimal user interaction. Unlike past methods, we do not make any restrictive assumptions about the geometry or texture of the object. To do so, we exploit the geometric and photometric similarity along and across the boundaries of matching fragments, and leverage user feedback to tackle the otherwise ill-posed problem. We begin by encoding the scale variability of each fragment's boundary contour in a multi-channel, 2D representation. Using this multi-channel boundary contour representation, we identify matching sub-contours via 2D partial image alignment. We then align the fragments by minimizing the distance between their adjoining regions while simultaneously ensuring geometric continuity across them. The configuration of the fragments as they are incrementally matched and aligned form a graph structure. By detecting cycles in this graph, we identify subsets of fragments with dependent alignments. We then minimize the error within the subsets to achieve a globally optimal alignment. Using ceramic pottery as the driving example, we demonstrate the accuracy and efficiency of our method on six real-world datasets.