We introduce a novel method for reconstructing the 3D geometry of botanical trees from
single photographs. Faithfully reconstructing a tree from single-view sensor data is a
challenging and open problem because many possible 3D trees exist that fit the tree's shape
observed from a single view. We address this challenge by defining a reconstruction
pipeline based on three neural networks. The networks simultaneously mask out trees in
input photographs, identify a tree's species, and obtain its 3D radial bounding volume-our …

Due to the dominant presence of vegetation, not only in outdoor but also in urban and even
indoor settings, detailed plant models significantly add to the realism of almost all virtual
scenes. This spans as wide as from applications in architecture and agriculture, research in
forestry and urban planning, training of autonomous agents, scene understanding for
augmented reality, and content creation for games and movies. It is important to reconstruct
models as faithful as possible to the captured plant's features, especially in settings where it …