Technology Category
Computer Vision: SegmentationConstraint Satisfaction and Optimization: Solvers and ToolsSearch and Optimization: Non-convex Optimization
Solving Convex Partition Visual Jigsaw Puzzles
🤖 AI Summary
This paper addresses the convex polygonal jigsaw puzzle problem—reconstructing an image from unordered, arbitrarily convex-shaped fragments to recover both their original geometric poses and semantic integrity. Unlike mainstream square-tile puzzle research, our approach removes restrictive shape assumptions and introduces a greedy solver that jointly optimizes geometric compatibility (edge lengths, angles, topological constraints) and visual content consistency (feature matching, texture continuity). We present CPuzzle, the first benchmark dataset specifically designed for convex-segmentation puzzles. Extensive experiments demonstrate that our method significantly outperforms existing baselines in reconstruction accuracy, robustness to noise and occlusion, and computational efficiency. This work extends the applicability of automated jigsaw solving beyond regular grids, establishing a new paradigm for irregular image restoration tasks—including archaeological artifact reconstruction, document forgery detection, and damaged imagery recovery.
Technology Category
Application Category
📝 Abstract
Jigsaw puzzle solving requires the rearrangement of unordered pieces into their original pose in order to reconstruct a coherent whole, often an image, and is known to be an intractable problem. While the possible impact of automatic puzzle solvers can be disruptive in various application domains, most of the literature has focused on developing solvers for square jigsaw puzzles, severely limiting their practical use. In this work, we significantly expand the types of puzzles handled computationally, focusing on what is known as Convex Partitions, a major subset of polygonal puzzles whose pieces are convex. We utilize both geometrical and pictorial compatibilities, introduce a greedy solver, and report several performance measures next to the first benchmark dataset of such puzzles.
Problem
Research questions and friction points this paper is trying to address.
Solving convex partition jigsaw puzzles computationally
Expanding puzzle types beyond traditional square pieces
Combining geometric and pictorial compatibilities for reconstruction
Innovation
Methods, ideas, or system contributions that make the work stand out.
Uses geometrical and pictorial compatibilities for solving
Introduces a greedy solver for convex partition puzzles
Creates benchmark dataset for polygonal jigsaw puzzles
🔎 Similar Papers
No similar papers found.
