Discriminative Learning of Markov Random Fields for Segmentation of 3D Scan Data
D. Anguelov, B.Taskar, V. Chatalbashev, D. Gupta, D. Koller, G. Heitz, A. Ng [CVPR 2005]


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.

Paper: [2.1MB PDF]
Flythrough movie: [3.1MB MP4 format] You will need Quicktime player to view this movie.

Below you can get supplementary materials for the datasets we experimented with in the paper.
Dataset 1: Terrain Classification
Dataset 2: Segmentation of Articulated Objects
Dataset 3: Object Segmentation for the Princeton Benchmark