GANzzle: Reframing jigsaw puzzle solving as a retrieval task using a generative mental image

TL;DR

This paper introduces GANzzle, a novel approach that frames jigsaw puzzle solving as a retrieval task using generative models to reconstruct images from unordered pieces, enabling size-agnostic puzzle solving.

Contribution

GANzzle leverages generative adversarial methods and a joint embedding space to solve puzzles as a retrieval problem, generalizing across multiple puzzle sizes.

Findings

Achieves puzzle size agnosticism unlike prior methods.

Performs on par with deep learning approaches on large-scale datasets.

Generalizes effectively to multiple puzzle sizes.

Abstract

Puzzle solving is a combinatorial challenge due to the difficulty of matching adjacent pieces. Instead, we infer a mental image from all pieces, which a given piece can then be matched against avoiding the combinatorial explosion. Exploiting advancements in Generative Adversarial methods, we learn how to reconstruct the image given a set of unordered pieces, allowing the model to learn a joint embedding space to match an encoding of each piece to the cropped layer of the generator. Therefore we frame the problem as a R@1 retrieval task, and then solve the linear assignment using differentiable Hungarian attention, making the process end-to-end. In doing so our model is puzzle size agnostic, in contrast to prior deep learning methods which are single size. We evaluate on two new large-scale datasets, where our model is on par with deep learning methods, while generalizing to multiple puzzle sizes.