Auxilliary Tool: A data structure that keeps track of how a classifier (specifically a cascaded classifier) is performing. This is useful for evaluating candidate features to add.
More...
#include <StructTypes.h>
Public Attributes | |
| double | chisq |
| Chi-Squared Error of Classifier evaluated on training set. Lower is better. | |
| int | pos_rejects |
| Number of positive examples rejected by classifier -- lower is better. | |
| int | neg_rejects |
| Number of negative examples rejected by classifier -- higher is better. | |
| int | total_pos |
| Total size of positive example pool. | |
| int | total_neg |
| Total size of negative example pool. | |
| int | prev_pos_rejects |
| Number of positive examples rejected by the classifier before, i.e. independent of the feature under consideration. | |
| int | prev_neg_rejects |
| Number of negative examples rejected by the classifier before, i.e. independent of the feature under consideration. | |
| double | time_per_patch |
| Time required to evaluate a patch with this feature. Lower is better. | |
| double | threshold |
| Rejection criterion used to generate these performance results. | |
Auxilliary Tool: A data structure that keeps track of how a classifier (specifically a cascaded classifier) is performing. This is useful for evaluating candidate features to add.
1.7.2