Density‐free test for symmetry verification in images

• DOI: 10.1002/asmb.2321
• Publication Date: 2018-03-07T06:25:03Z
• Authors: Agata Migalska
• Keywords: Reflection symmetry; Statistic

Presence of symmetry is utilized in multiple machine vision systems to help achieve their goals. In numerous scenarios, this goal verify that certain indeed exhibited by an image. However, we find there a shortage methods for verification would be capable asserting arbitrary reflectional or rotational symmetry. Using detectors merely perform improvident and not justified. We thus propose novel statistical test fulfills the requirement versatility. The proposed based on principle if image invariant some hypothesized transformation, then averaged image, obtained averaging pixel intensities input its transformed copy, looks exactly same as Adopting viewpoint images are visual messages convey information allows us expect amount both same. On contrary, incorrectly chosen transformation shall result content being different. Based equality, construct statistic show that, when samples large, asymptotically normally distributed. Finally, validity performance method, meticulous experimental study was performed large set images. results confirmed postulated ability method demonstrated with several examples.