Vision and Mobile Robotics Laboratory | Software
|
|
Home | Members | Projects | Publications | Software | Videos | Internal |
CalcStackDim takes as input a spin-image stack and its eigen-spin-images. It outputs the number of eigen-spin-images needed to reconstruct the spin-images in the stack to a specified level (based on correlation coefficient) using projections onto significant eigen-spin-images.
CalcStackDim.c contains all of the functionality for creating and displaying spin-image stacks generated using CalcStackDim. It handles mouse and keyboard events and spin-image animation.
By typing 'CalcStackDim -' the following options (format description default) are printed:
Usage: CalcStackDim (see CalcStackDim for complete usage)
%S input stack filename [required]
%S input eigen-stack filename [required]
%F reconstruction level [required]
-spinMethod %d method of spin image generation
%S input stack filename [required]
The spin-image stack from which the eigen-stack was computed.
%S input eigen-stack filename [required]
The eigen-stack for the input spin-images.
%F reconstruction level [required]
The level of reconstruction desired. For example, a level of reconstruction of 0.90 requires that at least half of the spin-images when correlated with their reconstructed images (created by projection onto the significant eigen-spin-images) will have a correlation coefficient of atleast 0.90.
-spinMethod %d method of spin image generation
Spin image generation method: 0=vertices only, 1=interpolate in spin-image space. Using method 1 allows matching and recognition between meshes of differing resolution, since they create spin-images based on the entier surface of the mesh rather than just the vertices . Method 1 projects each mesh face into spin-image space (the alpha-beta coordinate system) and then raster scans the projected shape to find the contribution of the face to the spin-image. This idea is described in Toward a General 3-D Matching Engine: Multiple Models, Complex Scenes, and Efficient Data Filtering.
To determine the number of significant eigen-spin-images for a duck model at a level of 0.95 use
CalcStackDim duckie.stack duckie.estack 0.95
The result printed will be something like:
Need 5 eigen-spin-images to reconstruct to 0.95 median correlation coefficient.