We address the problem of unsupervised learning of complex articulated
object models from 3D range data. We describe an algorithm whose input is a
set of meshes corresponding to different configurations of an articulated
object. The algorithm automatically recovers a decomposition of the object
into approximately rigid parts, the location of the parts in the di erent
object instances, and the articulated object skeleton linking the parts. Our
algorithm first registers all the meshes using an unsupervised non-rigid
technique described in a companion paper. It then segments the meshes using
a graphical model that captures the spatial contiguity of parts. The
segmentation is done using the EM algorithm, iterating between finding a
decomposition of the object into rigid parts, and finding the location of the
parts in the object instances. Although the graphical model is densely
connected, the object decomposition step can be performed optimally and
efficiently, allowing us to identify a large number of object parts while
avoiding local maxima. We demonstrate the algorithm on real world datasets,
recovering a 15-part articulated model of a human puppet from just 7
different puppet con gurations, as well as a 4 part model of a existing arm
where significant non-rigid deformation was present.