Abstract
A method that based on two-dimensional (2D) video sequences and employs the multilevel decomposition of wavelets and visualizes three-dimensional (3D) image video sequences in real-time mode is proposed. The method employs reconstructions of images with subsequent 3D visualization and includes the separation of a color image over RGB components. The NNI method is applied to perform the interpolation between the current and more precise estimates for a depth map at each decomposition level in the course of the motion from a rough estimate to a more precise one. The conventional L&K and SSD algorithms that show a worse quality of reconstructing three dimensional video sequences require 22.59 and 16.26 s, respectively, for the 3D-framevisualization. Using lenses equipped with red and blue optic filters for the right and left eyes, respectively, 3D video sequences are observed.
Original language | English |
---|---|
Pages (from-to) | 479-482 |
Number of pages | 4 |
Journal | Doklady Physics |
Volume | 56 |
Issue number | 9 |
DOIs | |
State | Published - Sep 2011 |