Quasi-Dense Reconstruction from Image Sequence

Abstract. This paper proposes a quasi-dense reconstruction from uncalibrated sequence. The main innovation is that all geometry is computed based on re-sampled quasi-dense correspondences rather than the standard sparse points of interest. It not only produces more accurate and robust reconstruction due to highly redundant and well spread input data, but also fills the gap of insufficiency of sparse reconstruction for visualization application. The computational engine is the quasi-dense 2-view and the quasi-dense 3-view algorithms developed in this paper. Experiments on real sequences demonstrate the superior performance of quasi-dense w.r.t. sparse reconstruction both in accuracy and robustness.

• Type: Conference
• Year: 2002
• Where: ECCV
• Authors: Maxime Lhuillier, Long Quan
• Keywords: Computer Vision, ECCV 2002, Quasi-dense 3-view Algorithms, Quasi-dense Reconstruction, Robust Reconstruction, Sparse Reconstruction, Standard Sparse Points