Optimal Representation of Multiple View Video
In Proceedings British Machine Vision Conference 2014
AbstractMultiple view video acquisition is widely used for reconstruction and free-viewpoint rendering of dynamic scenes by directly resampling from the captured images. This paper addresses the problem of optimally resampling and representing multiple view video to obtain a compact representation without loss of the view-dependent dynamic surface appearance. Spatio-temporal optimisation of the multiple view resampling is introduced to extract a coherent multi-layer texture map video. This resampling is combined with a surface-based optical flow alignment between views to correct for errors in geometric reconstruction and camera calibration which result in blurring and ghosting artefacts. The multi-view alignment and optimised resampling results in a compact representation with minimal loss of information allowing high-quality free-viewpoint rendering. Evaluation is performed on multiple view datasets for dynamic sequences of cloth, faces and people. The representation achieves >90% compression without significant loss of visual quality.
SessionVideo and Structure From Motion
FilesExtended Abstract (PDF, 1 page, 2.1M)
Paper (PDF, 12 pages, 3.1M)
Supplemental Materials (ZIP, 7.6M)