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

Talon, Davide; Del Bue, Alessio; James, Stuart

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.

Citation

Talon, D., Del Bue, A., & James, S. (2022, October). GANzzle: Reframing jigsaw puzzle solving as a retrieval task using generative mental images. Presented at IEEE International Conference on Image Processing, Bordeaux, France

• Presentation Conference Type: Conference Paper (published)
• Conference Name: IEEE International Conference on Image Processing
• Start Date: Oct 16, 2022
• Publication Date: 2022
• Deposit Date: Oct 24, 2024
• Peer Reviewed: Peer Reviewed
• Book Title: 2022 IEEE International Conference on Image Processing (ICIP)
• DOI: https://doi.org/10.1109/ICIP46576.2022.9897553
• Keywords: own, conference
• Public URL: https://durham-repository.worktribe.com/output/2024604