We propose a novel and robust model to represent and learn generic 3D object categories. We aim to solve the problem of true 3D object categorization for handling arbitrary rotati...
Abstract. We present a method for 3D object modeling and recognition which is robust to scale and illumination changes, and to viewpoint variations. The object model is derived fro...
In this paper we propose an algorithm to partition MPEG-4 video objects into temporal segments with uniform activity levels. These segments can be used for efficient summarization...
— It is frequently accepted in the manipulation literature that tactile sensing is needed to improve the precision of robot manipulation. However, there is no consensus on how th...
The ability to accurately localize objects in an observed scene is regarded as an important precondition for many practical applications including automatic manufacturing, quality ...