Shape registration with directional data. (July 2018)

• Record Type: Journal Article
• Authors: Grogan, Mairéad; Dahyot, Rozenn
• Is Part Of: Pattern recognition. Volume 79(2018:Jul.)
• Page Start: 452
• Page End: 466
• Publication Date: 2018-07
• Subjects: Shape registration -- Directional information -- Von Mises-Fisher -- L2 registration; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2018.02.021
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: We extend the L2 registration framework to include directional data. We model the directional data using von Mises Fisher kernels. We propose 4 cost functions for shape registration based on this modelling. Our method compares well with State of the Art techniques. Abstract: We propose several cost functions for registration of shapes encoded with Euclidean and/or non-Euclidean information (unit vectors). Our framework is assessed for estimation of both rigid and non-rigid transformations between the target and model shapes corresponding to 2D contours and 3D surfaces. The experimental results obtained confirm that using the combination of a point's position and unit normal vector in a cost function can enhance the registration results compared to state of the art methods.