Robust techniques for object recognition and pose estimation are essential for robotic manipulation and object grasping. In this paper, a novel approach for object recognition and pose estimation based on color cooccurrence histograms and geometric modelling is presented. The particular problems addressed are: (i) robust recognition of objects in natural scenes, (ii) estimation of partial pose ...

The aim of visual recognition is to identify objects in a scene and estimate their pose. Object recognition from 2D images is sensitive to illumination, pose, clutter and occlusions. Object recognition from range data on the other hand does not suffer from these limitations. An important paradigm of recognition is model-based whereby 3D models of objects are constructed offline and saved in a d...

This paper presents a method for object recognition, novel object detection, and estimation of the most salient object within a set. Objects are sampled using a scale invariant region detector, and each region is characterized by the subset of texture and color descriptors selected by a Genetic Algorithm (GA). Using multiple views of an object, and multiple regions per view, objects are modeled...

A novel active-vision system is presented in this paper for moving-object recognition. The proposed system is a natura! extension of the static-object recognition system developed earlier in our laboratory. Pre-marking, via circular markers placed on the objects, and active detection, via a camera mounted on a robot, comprise the two principles of the recognition technique. The primary steps of...

A theory of object recognition requires a theory of shape. Despite considerable empirical and theoretical research, however, a definition of object shape has proved elusive. Two experiments provide new insights by showing that children's object recognition changes dramatically during the period between 17 and 25 months. During this time, children develop the ability to recognize stylized three-...

This paper describes a novel view-based algorithm for 3D object recognition using a network of linear units. The SNoW learning architecture is a sparse network of linear functions over a pre-defined or incrementally learned feature space and is specifically tailored for learning in the presence of a very large number of features. We use the pixel-level representation in the experiments and comp...