GARF is a machine-learning system that can assemble 3D scans of broken objects to look like the original object, even if fragments are missing or edges are imperfect.
The objects could include bones, archaeological artefacts or other items, with the product of the programme being a 3D model.
While the system could theoretically be brute-forced, the intricacy of the work means machine learning is a much more practical solution.
The algorithm addresses the challenge of training on synthetic data and applying it to real-world breakage.
The GitHub repository contains the code, and a demo is available for those interested.
While the technology is new, cultural touchstones such as Batman and novel Snow Crash have explored the idea in popular culture.
