Drago Anguelov - Research Projects
Contact Info Projects

3D Modeling

SCAPE: Shape Completion and Animation of People

D. Anguelov, P. Srinivasan, D. Koller, S. Thrun, J. Rodgers, J. Davis
A data-driven approach for building a human shape model which spans variation in both subject shape and pose from 3D scans. The model is useful for a variety of animation and shape completion tasks. We can synthesize complete 3D surfaces for a subject using the output of a marker-based motion capture system. We can also use our model to complete a partial scans of different people in different poses.
[Go to Project Page]

The Correlated Correspondence Algorithm for Surface Registration

D. Anguelov, P. Srinivasan, D. Koller, S. Thrun, H.-C. Pang, J. Davis
An algorithm for registering 3D surface scans of an object undergoing significant deformations.  The algorithm registers two meshes by optimizing a joint probabilistic model over all point-to-point correspondences between them, which attempts to capture local geometry and preserve geodesic distances. The algorithm does not need markers, nor does it assume prior knowledge about object shape, the dynamics of its deformation, or scan alignment (although such knowledge can be incorporated if it is available).
[NIPS 2004]
[Go to Project Page]

Recovering Articulated Object Models from 3D Range Data

D. Anguelov, D. Koller, H.-C. Pang, P. Srinivasan, S. Thrun
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 different object instances, and the
articulated object skeleton linking the parts. It assumes the correspondences between the scans are known (we use the Correlated Correspondence algorithm above to recover them).
[UAI 2004]
[Go to Project Page]

3D Scene Segmentation

Discriminative Learning of Markov Random Fields for Segmentation of 3D Range Data

D. Anguelov, B. Taskar, V. Chatalbashev, D. Koller, D. Gupta, G. Heitz, A. Ng

We address the problem of segmenting 3D scan data into objects or object classes. Our segmentation framework is based on a subclass of Markov Random Fields (MRFs) which support efficient graph-cut inference. The MRF models  incorporate a large set of diverse features and enforce the preference that adjacent scan points have the same classification label. We use a recently proposed maximum-margin framework to discriminatively train the model from a set of labeled scans; as a result we automatically learn the relative importance of the features for the segmentation task.
[CVPR 2005]
[Go to Project Page]