SB SDK 1.11
sb_t_stat_model Struct Reference

Statistics of a model. More...

#include <sb.h>

Collaboration diagram for sb_t_stat_model:

Data Fields

int tp
 Count of True Positive. More...
 
int tn
 Count of True Negative. More...
 
int fp
 Count of False Positive. More...
 
int fn
 Count of False Negative. More...
 
int op
 Count of Optional Positive. More...
 
int on
 Count of Optional Negative. More...
 
int out_of_roi
 Count of Out Of ROI. More...
 
int mod_disabled
 Count of samples with model disabled. More...
 
sb_t_range_flt tp_weight
 Range of classification weight of the True Positive samples.
 
sb_t_range_flt tn_weight
 Range of classification weight of the True Negative samples.
 
sb_t_range_flt fp_weight
 Range of classification weight of the False Positive samples.
 
sb_t_range_flt fn_weight
 Range of classification weight of the False Negative samples.
 
sb_t_range_flt op_weight
 Range of classification weight of the Optional Positive samples.
 
sb_t_range_flt on_weight
 Range of classification weight of the Optional Negative samples.
 
sb_t_range_flt or_weight
 Range of classification weight of the out of the ROI samples.
 
float precision
 precision More...
 
float recall
 recall More...
 
float accuracy
 accuracy More...
 
float specificity
 specificity More...
 
float score
 Level score. More...
 

Detailed Description

Statistics of a model.

Note
In multi-model case may happen that a FALSE NEGATIVE sample is associated to a FALSE POSITIVE sample. In this case the error weighs double in the statistics calculation.
See also
sb_t_stat.

Definition at line 10306 of file sb.h.

Field Documentation

◆ accuracy

float sb_t_stat_model::accuracy

accuracy

accuracy = (tp + tn) / (tp + tn + fp + fn)

Definition at line 10408 of file sb.h.

◆ fn

int sb_t_stat_model::fn

Count of False Negative.

Definition at line 10347 of file sb.h.

◆ fp

int sb_t_stat_model::fp

Count of False Positive.

  • Retina: count of False Positive samples.
  • Surface, Deep Surface: count of False Positive blobs if blob analysis has been performed. Otherwise the count refers to blocks.
  • Deep Cortex: count of False Positive images.

Definition at line 10337 of file sb.h.

◆ mod_disabled

int sb_t_stat_model::mod_disabled

Count of samples with model disabled.

Definition at line 10383 of file sb.h.

◆ on

int sb_t_stat_model::on

Count of Optional Negative.

  • Retina: count of Optional Negative samples.
  • Surface, Deep Surface: count of Optional Negative blobs if blob analysis has been performed. Otherwise the count refers to blocks.
  • Deep Cortex: not used.

Definition at line 10366 of file sb.h.

◆ op

int sb_t_stat_model::op

Count of Optional Positive.

Definition at line 10357 of file sb.h.

◆ out_of_roi

int sb_t_stat_model::out_of_roi

Count of Out Of ROI.

  • Retina: count of Out Of ROI samples.
  • Surface, Deep Surface: count of Out Of ROI blobs if blob analysis has been performed. Otherwise the count refers to blocks.
  • Deep Cortex: not used.

Definition at line 10375 of file sb.h.

◆ precision

float sb_t_stat_model::precision

precision

precision = tp / (tp + fp)

Definition at line 10396 of file sb.h.

◆ recall

float sb_t_stat_model::recall

recall

recall = tp / (tp + fn)

Definition at line 10402 of file sb.h.

◆ score

float sb_t_stat_model::score

Level score.

Score of the model on the processed level. It is computed only in SVL mode.
The range of the values is [0, 1]
Used only with Surface projects.

Definition at line 10422 of file sb.h.

◆ specificity

float sb_t_stat_model::specificity

specificity

specificity = tn / (tn + fp)

Definition at line 10414 of file sb.h.

◆ tn

int sb_t_stat_model::tn

Count of True Negative.

  • Retina: count of True Negative samples. Moreover, the value is set to 1 if an image has no samples and no occurrences have been found.
  • Surface, Deep Surface: count of True Negative blobs if blob analysis has been performed. Otherwise the count refers to blocks.
  • Deep Cortex: count of True Negative images.

Definition at line 10327 of file sb.h.

◆ tp

int sb_t_stat_model::tp

Count of True Positive.

  • Retina: count of True Positive samples.
  • Surface, Deep Surface: count of True Positive blobs if blob analysis has been performed. Otherwise the count refers to blocks.
  • Deep Cortex: count of True Positive images.

Definition at line 10316 of file sb.h.


The documentation for this struct was generated from the following file: