- CNR-IMATI Milano
Wednesday, December 20, 2017
There has been significant prior work on learning realistic, articulated, 3D statistical shape models of the human body. In contrast, there are few such models for animals, despite their many applications in biology, neuroscience, agriculture, and entertainment.
The main challenge is that animals are much less cooperative subjects than humans: the best human body models are learned from thousands of 3D scans of people in specific poses, which is infeasible with live animals.
In the talk I will illustrate our work where we extend a state-of-the-art articulated 3D human body model to animals and learn it from a limited set of 3D scans of toy figurines in arbitrary poses.
We employ a novel part-based shape model to compute an initial registration to the scans, we then normalize their pose, and learn a statistical shape model. In this way, we accurately align animal scans from different quadruped families with very different shapes and poses. With the alignment to a common template we learn a shape space representing animals including lions, cats, dogs, horses, cows and hippos.
Animal shapes can be sampled from the model, posed, animated, and fitted to data.
The generalization of the model is illustrated by fitting it to images of real animals, where it captures realistic animal shapes, even for new species not seen in training.
Contact person: Umberto Castellani
- Contact person
- Publication date
December 12, 2017