Friday 24 May 2013
 
» The Fiber Cup » How to evaluate your own results?
 

 
 

How to evaluate your own results?

 

You can get the source code of the program used for evaluation here. Binaries for the most standard systems are also available. Consider using them instead of recompiling the program from scratch.

The program ’fibercup’ contains three commands. By typing ’fibercup’ in a terminal, the list of available commands will show up:


>fibercup
Software The Fiber Cup (c) 2010 INRIA, version 1.0.0

Author: Pierre Fillard (Pierre.Fillard@inria.fr)

This program comes with ABSOLUTELY NO WARRANTY;
This is free software, and you are welcome to redistribute it
under certain conditions; See License.txt for more information.

Available commands:
        compare
        resample
        warp


The most important command is ’compare’. By typing ’fibercup compare’, a help message will be displayed:


>fibercup compare
Software The Fiber Cup (c) 2010 INRIA, version 1.0.0

Author: Pierre Fillard (Pierre.Fillard@inria.fr)

This program comes with ABSOLUTELY NO WARRANTY;
This is free software, and you are welcome to redistribute it
under certain conditions; See License.txt for more information.

Available commands:
Usage: compare
<-s: sampling> <-m: metric (l2/tan/curv)> <-q: quantity (mean/std/min/max)> <-t1: type (interp/approx)> <-t2: type (interp/approx)> <-i1: input 1> <-i2: input 1>

Compare two fibers using the L2, curve or tangent metric


Description of the options:

  • -s: specify the sampling of the fibers to perform evaluation. For the Fiber Cup, a sampling of 1000 was used for every fiber.
  • -m: specify which metric to use between the spatial metric (l2), the tangent metric (tan), and the curve metric (curv)
  • -q: which quantity should be calculated. ’mean’ is for the average value of the chosen metric estimated over the number of sampling points. ’std’ stands for the standard deviation of the same quantity. ’min’ (resp. ’max’) returns the minimum (resp. maximum) value of the chosen metric.
  • -t1: choose either interpolating of approximating splines for the input 1. In general, interpolation should be used not to alter the positions return by a competitor.
  • -t2: choose either interpolating of approximating splines for the input 2. In general, interpolation should be used not to alter the positions return by a competitor.
  • -i1: filename of input 1 (fiber text file)
  • -i2: filename of input 2 (fiber text file)


Example:


>fibercup compare -s 1000 -m l2 -q mean -t1 interp -t2 interp -i1 ground_truth_1.txt -i2 candidate_1.txt

Compute the average spatial metric using 1000 points between fibers ground_truth_1.txt and candidate_1.txt using interpolating splines.

The program prints out the requested quantity.

 
 
Published on Tuesday 29 June 2010

 
 
 
The others articles in the same section :
 
Published on Friday 24 July 2009
Update Thursday 30 July 2009
 
Published on Saturday 25 July 2009
Update Tuesday 1 February 2011
 
Published on Monday 20 July 2009
Update Tuesday 29 June 2010
 
Published on Thursday 26 February 2009
 
Published on Saturday 25 July 2009
Update Monday 27 July 2009
 
Published on Monday 20 July 2009
 
Published on Saturday 25 July 2009
 

 
Home     |    Site Map     |    Editors area     |    Login
  Update Monday 14 November 2011