Learning object classes with generic knowledge
Thomas Deselaers, Bogdan Alexe and Vittorio Ferrari
(2011) Technical Report. ETH Zurich, Zurich, Switzerland.

Abstract

Learning a new object class from cluttered training images is very challenging when the location of object instances is unknown, i.e. in a weakly su- pervised setting. Many previous works require objects covering a large portion of the images. We present a novel approach that can cope with extensive clutter as well as large scale and appearance variations be- tween object instances. To make this possible we ex- ploit generic knowledge learned beforehand from images of other classes for which location annotation is avail- able. Generic knowledge facilitates learning any new class from weakly supervised images, because it reduces the ambiguity in the location of its object instances. We propose a conditional random eld that starts from generic knowledge and then progressively adapts to the new class. Our approach simultaneously localizes object instances while learning an appearance model specic for the class. We demonstrate this on several datasets, including the very challenging Pascal VOC 2007. Fur- thermore, our method enables to train any state-of-the- art object detector in a weakly supervised fashion, al- though it would normally require object location anno- tations.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.

[Edit]