Dynamic face pose change and noise make it difficult to track multi-view faces in a cluttering environment. In this paper, we propose a graphical model based method, which combin...
When articulated figures interact in a 3D environment, collisions are highly likely and must often be avoided. We present a method automatically producing realistic collision-free...
City environments often lack textured areas, contain
repetitive structures, strong lighting changes and therefore
are very difficult for standard 3D modeling pipelines.
We prese...
This paper presents a method for visual object categorization based on encoding the joint textural information in objects and the surrounding background, and requiring no segmenta...
Alireza Tavakoli Targhi, Andrzej Pronobis, Heydar ...
We present a novel object-specific segmentation method which can be used in view-based object recognition systems. Previous object segmentation approaches generate inexact results ...
Minsu Cho (Seoul National University), Kyoung Mu L...