Creating 3D environments of paintings used for learning geometric concepts

Abstract

Recreating single 2D images to 3D environments is one of the classic computer vision problems. Specifcally recreating paintings introduces the problem of having an incorrect per-spective, which makes reconstruction with automatic and semi-automatic methods impossible. This means that the paintings have to be reconstructed manually using 3D modeling software. We present an application that contains a 3D environment to teach children about perspective. In order to teach perspective, tasks are created that are derived from assignments in the CITO exams. The tasks that children have to do, derived from the assignments, can be used in any reconstructed painting. One painting has been recreated and duplicated to show that task generation is random and works in any painting. The application shows there might be educational potential for learning about both perspective and the painting itself. However, the group of children that took part in the experiment is too small to make concrete statements about the results.