Reassembling Fragmented Objects through Geometric Breaking Curves Registration

Abstract

The field of archaeology is constantly faced with the challenge of restoring broken artifacts. Traditional manual restoration methods can be time-consuming and prone to errors, making it difficult to scale the process. As technology has advanced, many methods have been proposed to assist the restoration process. However, most of these methods target specific object types and operate under strict constraints. In this thesis, we propose a digital pairwise alignment framework as the first step for a full-shape restoration pipeline. The framework consists of robust, proven methods for automatic fragment reassembly of generic three-dimensional objects of any type. Our pipeline includes a novel segmentation approach that is able to handle the reassembly of objects from heavily damaged fragments by using both fracture surfaces and salient features on the intact sides of fragments when available. The pipeline only uses geometric cues of the objects, making the registration procedure easier and extendible to other object types. Our approach has been tested on both synthetic and real-world fractured objects acquired within the H2020 project RePAIR. Overall, our pipeline for pairwise fragment alignment offers a promising solution for the automatic reassembly of broken artifacts, which can save time and reduce errors in the restoration process.