Normal?Driven Spherical Shape Analogies

This paper introduces a new method to stylize 3D geometry. The key observation is that the surface normal an effective instrument capture different geometric styles. Centered around this observation, we cast stylization as shape analogy problem, where relationship defined on normal. formulation can deform into styles within single framework. One plug-and-play target by providing exemplar or energy-based style description (e.g., developable surfaces). Our methodology enables Normal Captures counterpart material captures (MatCaps) used in rendering, and prototypical concept of Spherical Shape Analogies image analogies processing.