RGB-D Based Detection of Texture-Less Planar Objects Applied to an AR Jigsaw Puzzle

This paper presents an Augmented Reality application that aims to help players to solve a jigsaw puzzle that consists of texture-less planar pieces with a discriminative shape. The pieces are detected using a technique based on RGB-D consumer devices, their poses are estimated, correctly assembled pieces are highlighted and additional information is rendered over each piece. A user study was conducted to evaluate the system, which demonstrated that players who used the application for solving the jigsaw puzzle obtained better results than the ones who did not use it.