Multi-scale 3D scene flow from binocular stereo sequences
Computer Vision and Image Understanding
Real-time joint disparity and disparity flow estimation on programmable graphics hardware
Computer Vision and Image Understanding
Illumination-robust variational optical flow using cross-correlation
Computer Vision and Image Understanding
Dense and accurate spatio-temporal multi-view stereovision
ACCV'09 Proceedings of the 9th Asian conference on Computer Vision - Volume Part II
Over-Parameterized optical flow using a stereoscopic constraint
SSVM'11 Proceedings of the Third international conference on Scale Space and Variational Methods in Computer Vision
| Hi-index | 0.00 |
We present a common variational framework for dense depth recoveryand dense three-dimensional motion field estimation from multiplevideo sequences, which is robustto camera spectral sensitivitydifferences and illumination changes. For this purpose, we firstshow that both problems reduce to a generic image matching problemafter backprojecting the input images onto suitable surfaces. Wethen solve this matching problem in the case of statisticalsimilarity criteria that can handle frequently occurring non-affineimage intensities dependencies. Our method leads to an efficientand elegant implementation based on fast recursive filters. Weobtain good results on real images.