SymmMap: Estimation of the 2-D Reflection Symmetry Map and Its Applications

Rajendra Nagar, Shanmuganathan Raman; Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017, pp. 1715-1724

Detecting the reflection symmetry axis present in an object has been an active research problem in computer vision and computer graphics due to its various applications such as object recognition, object detection, modelling, and symmetrization of 3D objects. However, the problem of computing the reflection symmetry map for a given image containing objects exhibiting reflection symmetry has received a very little attention. The symmetry map enables us to represent the pixels in the image using a score depending on the probability of each of them having a symmetric counterpart. In this work, we attempt to compute the 2-D reflection symmetry map. We pose the problem of generating the symmetry map as an intra-image dense symmetric pixels correspondence problem, which we solve efficiently using a randomized algorithm by observing the reflection symmetry coherency present in the image. We introduce an application of symmetry map called symmetry preserving image stylization.