ABSTRACT
Modeling the interesting and complex geometry of plants has been a center of attention of research for biologists and mathematicians for decades. Numerous approaches have been proposed in order to mathematically model the geometrical structure of plants in a robust manner. Whereas one motivation for studying plant geometry is to better understand the structure of plants from a mathematical perspective, realistic modeling of plants has become an active area of research in computer graphics, in particular, due to the explosion of plant phenotyping platforms. This chapter focuses on the state-of-the-art techniques used to reconstruct the plant geometry from real data (data-driven modeling techniques). The hybrid approach is based on a combined framework of procedural and data-driven modeling techniques. Zhang et al. proposed a technique by first constructing the visible parts of the scanned input point cloud data, and then synthesizing the non-visible parts, using a shape prediction model.
