2 Hits in 2.0 sec

GeoF: Geodesic Forests for Learning Coupled Predictors

Peter Kontschieder, Pushmeet Kohli, Jamie Shotton, Antonio Criminisi
2013 2013 IEEE Conference on Computer Vision and Pattern Recognition  
GeoF outperforms both stateof-the-art forest models and the conventional pairwise CRF.  ...  Conventional decision forest based methods for image labelling tasks like object segmentation make predictions for each variable (pixel) independently [3, 5, 8] .  ...  Wen for making available the labelled face images to us. We are also very grateful to S. Nowozin, C. Rother, A. Fitzgibbon and J. Jancsary for inspiring and heated conversations.  ... 
doi:10.1109/cvpr.2013.16 dblp:conf/cvpr/KontschiederKSC13 fatcat:ovuftgtg75gyjjb3xb6w34rk64

Deep Neural Decision Forests

Peter Kontschieder, Madalina Fiterau, Antonio Criminisi, Samuel Rota Bulo
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
GeoF: trees for classification and regression. (PAMI), 21(12):1297– Geodesic forests for learning coupled predictors.  ...  Learning Prediction Nodes updating individual trees instead of the entire forest reduces Given the update rules for the decision function parame- the computational  ... 
doi:10.1109/iccv.2015.172 dblp:conf/iccv/KontschiederFCB15 fatcat:xzvdxd6qxna3blchaadgkmtfxy