Abstract— Recently, classical pairwise Structure From Motion (SfM) techniques have been combined with non-linear global optimization (Bundle Adjustment, BA) over a sliding window...
Javier Civera, Oscar G. Grasa, Andrew J. Davison, ...
This paper describes an approach for video structuring and indexing. It relies on motion wavelet coefficients directly estimated from image sequence. These coefficients provide a ...
Extracting meaningful 3D human motion information from video sequences is of interest for applications like intelligent humancomputer interfaces, biometrics, video browsing and ind...
Abstract. In spite of over two decades of intense research, illumination and pose invariance remain prohibitively challenging aspects of face recognition for most practical applica...
— Obtaining models of dynamic 3D objects is an important part of content generation for computer graphics. Numerous methods have been extended from static scenarios to model dyna...