compute_statistical_measures

menpofit.error.compute_statistical_measures(errors, step_error, max_error, min_error=0.0)[source]

Computes various statistics given a set of errors that correspond to multiple images. It can also deal with multiple sets of errors that correspond to different methods.

Parameters:
  • errors (list of float or list of list of float) – The list of errors per image. You can provide a list of lists for the errors of multiple methods.
  • step_error (float) – The sampling step of the error bins of the CED for computing the Area Under the Curve and the Failure Rate.
  • max_error (float) – The maximum error value of the CED for computing the Area Under the Curve and the Failure Rate.
  • min_error (float) – The minimum error value of the CED for computing the Area Under the Curve and the Failure Rate.
Returns:

  • mean (float or list of float) – The mean value.
  • mean (float or list of float) – The standard deviation.
  • median (float or list of float) – The median value.
  • mad (float or list of float) – The mean absolute deviation value.
  • max (float or list of float) – The maximum value.
  • auc (float or list of float) – The area under the curve value.
  • fr (float or list of float) – The failure rate value.