We propose use of an appearance manifold with embedded covariance matrix as a technique for recognizing 3D objects from images that are influenced by geometric and quality-degrade...
Segmentation of range images using superquadric entities has been pointed out by a number of researchers as a powerful approach towards object recognition. Problems exist in findi...
In order to perform object recognition it is necessary to learn representations of the underlying components of images. Such components correspond to objects, object-parts, or fea...
Here, we present a constrained object recognition task that has been robustly solved largely with simple machine learning methods, using a small corpus of about 100 images taken u...
A fast simulatedannealingalgorithmis developed for automatic object recognition. The object recognition problem is addressed as the problem of best describing a match between a hy...