P000227R1
Robust Symmetry Detection for 2D Shapes based on Electrical Charge Distribution
*Zhiyang Li (College of Information Science and Technology, Dalian Maritime University)
Wenyu Qu (College of Information Science and Technology, Dalian Maritime University)
Yujie Xu (College of Information Science and Technology, Dalian Maritime University)
Heng Qi (School of Computer Science and Technology, Dalian University of Technology)
Junjie Cao (School of Mathematical Sciences, Dalian University of Technology)
Shape symmetry detection plays a crucial role in human recognition tasks. Although a variety of symmetry detection methods for two dimensional shape have been proposed in the literature, there is still a need for a robust and versatile method to detect varied forms of symmetries, such as extrinsic vs. intrinsic symmetry, and global vs. local symmetry. To address this issue, we present an effective symmetry detection method based on the electrical charge distribution on the shape (ECDS). The motivation is that electrical charge distribution will exhibit some symmetries if the shape possesses symmetries. Upon obtaining ECDS, the problem of detecting symmetry in the shape is transformed as a problem of detecting horizontal and vertical lines in a local similarity matrix. This strategy make our algorithm more flexible and feasible. The algorithm is able to detect not only the extrinsic and intrinsic symmetry, but also all the local reflectional symmetry. Extensive experiments have been done on both partially symmetric and symmetric shape, demonstrating the robustness and effectiveness of our symmetry detection method.