Structure-from-Sherds++: Robust Incremental 3D Reassembly of Axially Symmetric Pots from Unordered and Mixed Fragment Collections

TL;DR

This paper introduces Structure-from-Sherds++, a novel incremental 3D reassembly method for axially symmetric pots that effectively filters false matches and reconstructs multiple objects simultaneously, outperforming existing approaches.

Contribution

The paper presents a new iterative registration approach with multi-graph beam search for robust, scalable 3D reassembly of mixed fragment collections without prior object information.

Findings

Achieves 87% reassembly accuracy on real fragment dataset.

Outperforms existing methods in handling complex fracture patterns.

Effectively reconstructs multiple pots simultaneously.

Abstract

Reassembling multiple axially symmetric pots from fragmentary sherds is crucial for cultural heritage preservation, yet it poses significant challenges due to thin and sharp fracture surfaces that generate numerous false positive matches and hinder large-scale puzzle solving. Existing global approaches, which optimize all potential fragment pairs simultaneously or data-driven models, are prone to local minima and face scalability issues when multiple pots are intermixed. Motivated by Structure-from-Motion (SfM) for 3D reconstruction from multiple images, we propose an efficient reassembly method for axially symmetric pots based on iterative registration of one sherd at a time, called Structure-from-Sherds++ (SfS++). Our method extends beyond simple replication of incremental SfM and leverages multi-graph beam search to explore multiple registration paths. This allows us to effectively filter out indistinguishable false matches and simultaneously reconstruct multiple pots without requiring prior information such as base or the number of mixed objects. Our approach achieves 87% reassembly accuracy on a dataset of 142 real fragments from 10 different pots, outperforming other methods in handling complex fracture patterns with mixed datasets and achieving state-of-the-art performance. Code and results can be found in our project page https://sj-yoo.info/sfs/.