Kaye, DavidIvrissimtzis, IonnisJohn Collomosse and Ian Grimstead2014-01-312014-01-312010978-3-905673-75-3https://doi.org/10.2312/LocalChapterEvents/TPCG/TPCG10/127-130We propose a new algorithm for implicit surface reconstruction and feature detection. The algorithm is based on a self organising map with the connectivity of a regular 3D grid that can be trained into an implicit representation of surface data. The implemented self organising map stores not only its current state but also its recent training history which can be used for feature detection. Preliminary results show that the proposed algorithm gives good quality reconstructions and can detect various types of feature.Categories and Subject Descriptors (according to ACM CCS): I.3.5 [Computer Graphics]: Computational Geometry and Object Modeling I.6.5 [Simulation and modeling]: Model Development