We propose a new image registration criterion that uses segmented versions of the images to exploit both the geometric and photometric characteristics of the distinct regions they...
One of the original goals of computer vision was to fully understand a natural scene. This requires solving several sub-problems simultaneously, including object detection, region...
Geremy Heitz, Stephen Gould, Ashutosh Saxena, Daph...
A methodology for the segmentation of successive frames of a video sequence is presented. Traditional methods, treating each frame in isolation, are computationally expensive, ign...
: The basic line of our action is first to use natural language processing to prune the texts and the query, and secondly to use an ontology to expand the queries. Last year The sy...
While great strides have been made in detecting and localizing specific objects in natural images, the bottom-up segmentation of unknown, generic objects remains a difficult chall...