Reflexive symmetry detection in single image

Abstract

Reflective symmetry can be used as a strong prior for many computer vision tasks. We interpret the planar reflective symmetry detection by using the property of an involution, which implies that two pairs of matched points are enough to define a planar reflective symmetry observed from a non-frontal viewpoint. This makes the reflective symmetry estimation as efficient as the classical homography estimation in binocular stereovision. This simple reflective symmetry computation can be plugged into any multiple model estimation to detect multiple symmetries at different scales and locations in images. The experimental results show that the proposed method is able to detect single and multiple reflective symmetries both in frontal and non fronto-parallel viewpoints.