Timezone: »

Spatial distance dependent Chinese Restaurant Process for image segmentation
Soumya Ghosh · Andrei B Ungureanu · Erik Sudderth · David Blei

Mon Dec 12 10:00 AM -- 02:59 PM (PST) @

The distance dependent Chinese restaurant process (ddCRP) was recently introduced to accommodate random partitions of non-exchangeable data. The ddCRP clusters data in a biased way: each data point is more likely to be clustered with other data that are near it in an external sense. This paper examines the ddCRP in a spatial setting with the goal of natural image segmentation. We explore the biases of the spatial ddCRP model and propose a novel hierarchical extension better suited for producing ""human-like"" segmentations. We then study the sensitivity of the models to various distance and appearance hyperparameters, and provide the first rigorous comparison of nonparametric Bayesian models in the image segmentation domain. On unsupervised image segmentation, we demonstrate that similar performance to existing nonparametric Bayesian models is possible with substantially simpler models and algorithms.

Author Information

Soumya Ghosh (MIT-IBM Watson AI Lab, IBM Research)
Andrei B Ungureanu (PDT Partners)
Erik Sudderth (University of California, Irvine)
David Blei (Columbia University)

More from the Same Authors