000171878 001__ 171878
000171878 005__ 20180913060951.0
000171878 020__ $$a978-3-642-03797-9
000171878 02470 $$2ISI$$a000279102000016
000171878 037__ $$aCONF
000171878 245__ $$aTraining for Task Specific Keypoint Detection
000171878 269__ $$a2009
000171878 260__ $$bSpringer-Verlag New York, Ms Ingrid Cunningham, 175 Fifth Ave, New York, Ny 10010 Usa$$c2009
000171878 336__ $$aConference Papers
000171878 490__ $$aLecture Notes in Computer Science
000171878 520__ $$aIn this paper, we show that a better performance can be achieved by training a keypoint detector to only find those points that are suitable to the needs of the given task. We demonstrate our approach in an urban environment, where the keypoint detector should focus on stable man-made structures and ignore objects that undergo natural changes such as vegetation and clouds. We use Wald-Boost learning with task specific training samples in order to train a keypoint detector with this capability. We show that our aproach generalizes to a broad class of problems where the task is known beforehand.
000171878 6531_ $$aFeatures
000171878 700__ $$0244088$$aStrecha, Christoph$$g182325
000171878 700__ $$aLindner, Albrecht
000171878 700__ $$0242723$$aAli, Karim$$g179297
000171878 700__ $$0240252$$aFua, Pascal$$g112366
000171878 7112_ $$aDAGM Symposium on Pattern Recognition$$cJena, GERMANY$$dSep 09-11, 2009
000171878 773__ $$j5748$$q151-160$$tPattern Recognition, Proceedings
000171878 8564_ $$s409557$$uhttps://infoscience.epfl.ch/record/171878/files/final.pdf$$yPreprint$$zPreprint
000171878 909C0 $$0252087$$pCVLAB$$xU10659
000171878 909CO $$ooai:infoscience.tind.io:171878$$pconf$$pIC
000171878 917Z8 $$x112366
000171878 937__ $$aEPFL-CONF-171878
000171878 973__ $$aEPFL$$rNON-REVIEWED$$sPUBLISHED
000171878 980__ $$aCONF