Showing 1–1 of 1 results for author: Hallman, S

Oriented Edge Forests for Boundary Detection

Authors: Sam Hallman, Charless C. Fowlkes

Abstract: We present a simple, efficient model for learning boundary detection based on a random forest classifier. Our approach combines (1) efficient clustering of training examples based on simple partitioning of the space of local edge orientations and (2) scale-dependent calibration of individual tree output probabilities prior to multiscale combination. The resulting model outperforms published result… ▽ More We present a simple, efficient model for learning boundary detection based on a random forest classifier. Our approach combines (1) efficient clustering of training examples based on simple partitioning of the space of local edge orientations and (2) scale-dependent calibration of individual tree output probabilities prior to multiscale combination. The resulting model outperforms published results on the challenging BSDS500 boundary detection benchmark. Further, on large datasets our model requires substantially less memory for training and speeds up training time by a factor of 10 over the structured forest model. △ Less

Submitted 28 June, 2015; v1 submitted 12 December, 2014; originally announced December 2014.

Comments: updated to include contents of CVPR version + new figure showing example segmentation results