Perceptual Models of Preference in 3D Printing Direction

摘要

This paper introduces a perceptual model for determining 3D printingrnorientations. Additive manufacturing methods involving lowcostrn3D printers often require robust branching support structuresrnto prevent material collapse at overhangs. Although the designedrnshape can successfully be made by adding supports, residual materialrnremains at the contact points after the supports have been removed,rnresulting in unsightly surface artifacts. Moreover, fine surfacerndetails on the fabricated model can easily be damaged whilernremoving supports. To prevent the visual impact of these artifacts,rnwe present a method to find printing directions that avoid placingrnsupports in perceptually significant regions. Our model for preferencernin 3D printing direction is formulated as a combination of metricsrnincluding area of support, visual saliency, preferred viewpointrnand smoothness preservation. We develop a training-and-learningrnmethodology to obtain a closed-form solution for our perceptualrnmodel and perform a large-scale study. We demonstrate the performancernof this perceptual model on both natural and man-madernobjects.