This algorithm is a legacy one. The API has changed since its implementation. New versions and forks will need to be updated.

This algorithm is an analyzer. It can only be used on analysis blocks.

Endpoint Groups 2

Algorithms have at least one input and one output. All algorithm endpoints are organized in groups. Groups are used by the platform to indicate which inputs and outputs are synchronized together. The first group is automatically synchronized with the channel defined by the block in which the algorithm is deployed.

Group: test

Endpoint Name	Data Format	Nature
test_fusion_scores	pkorshunov/fusion_scores/1	Input

Group: dev

Endpoint Name	Data Format	Nature
dev_fusion_scores	pkorshunov/fusion_scores/1	Input

Results 26

Analyzers may produce any number of results. Once experiments using this analyzer are done, you may display the results or filter experiments using criteria based on them.

Name	Type
number_of_positives_dev_spoof	int32
number_of_negatives_dev_spoof	int32
number_of_positives_test_spoof	int32
HTER_licit	float32
FRR_test_spoof	float32
roc_test	plot/isoroc/1
number_of_negatives_test_licit	int32
FAR_dev_spoof	float32
number_of_negatives_dev_licit	int32
FAR_dev_licit	float32
dev_histograms	plot/bar/1
number_of_positives_dev_licit	int32
EER_spoof	float32
test_histograms	plot/bar/1
HTER_spoof	float32
FAR_test_spoof	float32
number_of_positives_test_licit	int32
FRR_dev_licit	float32
number_of_negatives_test_spoof	int32
FRR_test_licit	float32
threshold_licit	float32
roc_dev	plot/isoroc/1
threshold_spoof	float32
FAR_test_licit	float32
FRR_dev_spoof	float32
EER_licit	float32

Parameters 3

Parameters allow users to change the configuration of an algorithm when scheduling an experiment

Name	Description	Type	Default
n_bins_real	Number of histogram bins for real scores	uint32	20
n_bins_attack	Number of histogram bins for attack scores	uint32	80
n_bins_zimp	Number of histogram bins for zero-impostors scores	uint32	100

import bob.measure
import numpy

class Algorithm:

def __init__(self):
        self.positives_dev_licit = []
        self.negatives_dev_licit = []
        self.positives_test_licit = []
        self.negatives_test_licit = []
        self.positives_dev_spoof = []
        self.negatives_dev_spoof = []
        self.positives_test_spoof = []
        self.negatives_test_spoof = []
        self.dev_scores_real = None
        self.n_bins_real = 20
        self.n_bins_zimp = 100
        self.n_bins_attack = 80

def setup(self, parameters):
        self.n_bins_real = float(parameters.get('n_bins_real', self.n_bins_real))
        self.n_bins_zimp = float(parameters.get('n_bins_zimp', self.n_bins_zimp))
        self.n_bins_attack = float(parameters.get('n_bins_attack', self.n_bins_attack))
        return True

def process(self, inputs, output):
        dev_scores_zimp = []
        dev_scores_attack = []
        test_scores_real = []
        test_scores_zimp = []
        test_scores_attack = []
        
        # accumulate the dev scores
        if self.dev_scores_real is None:
            self.dev_scores_real = []
            while inputs['dev_fusion_scores'].hasMoreData():
                inputs['dev_fusion_scores'].next()
                dev_fusion_scores = inputs['dev_fusion_scores'].data
                self.dev_scores_real.extend(dev_fusion_scores.real_scores)
                dev_scores_zimp.extend(dev_fusion_scores.zimp_scores)
                dev_scores_attack.extend(dev_fusion_scores.attack_scores)
          
        test_fusion_scores = inputs['test_fusion_scores'].data
        test_scores_real.extend(test_fusion_scores.real_scores)
        test_scores_zimp.extend(test_fusion_scores.zimp_scores)
        test_scores_attack.extend(test_fusion_scores.attack_scores)

# once all values are received, evaluate the scores
        if not(inputs.hasMoreData()):
            # first, compute everything for licit scenario (no attacks)
            self.positives_dev_licit = self.dev_scores_real
            self.negatives_dev_licit = dev_scores_zimp
            self.positives_test_licit = test_scores_real
            self.negatives_test_licit = test_scores_zimp
            
            eer_threshold_licit = bob.measure.eer_threshold(self.negatives_dev_licit, self.positives_dev_licit)
            far_dev_licit, frr_dev_licit   = bob.measure.farfrr(self.negatives_dev_licit, self.positives_dev_licit, eer_threshold_licit)
            far_test_licit, frr_test_licit = bob.measure.farfrr(self.negatives_test_licit, self.positives_test_licit, eer_threshold_licit)
            roc_points_dev_licit     = bob.measure.roc(self.negatives_dev_licit, self.positives_dev_licit, 100)
            roc_points_test_licit    = bob.measure.roc(self.negatives_test_licit, self.positives_test_licit, 100)

dev_hist_pos_licit, dev_bin_pos_licit = numpy.histogram(self.positives_dev_licit, self.n_bins_real)
            dev_hist_neg_licit, dev_bin_neg_licit = numpy.histogram(self.negatives_dev_licit, self.n_bins_zimp)
            test_hist_pos_licit, test_bin_pos_licit = numpy.histogram(self.positives_test_licit, self.n_bins_real)
            test_hist_neg_licit, test_bin_neg_licit = numpy.histogram(self.negatives_test_licit, self.n_bins_zimp)
            
            # then, compute everything for spoof scenario (no zero-impostor)
            self.positives_dev_spoof = self.dev_scores_real
            self.negatives_dev_spoof = dev_scores_attack
            self.positives_test_spoof = test_scores_real
            self.negatives_test_spoof = test_scores_attack
            eer_threshold_spoof = bob.measure.eer_threshold(self.negatives_dev_spoof, self.positives_dev_spoof)
            far_dev_spoof, frr_dev_spoof   = bob.measure.farfrr(self.negatives_dev_spoof, self.positives_dev_spoof, eer_threshold_spoof)
            far_test_spoof, frr_test_spoof = bob.measure.farfrr(self.negatives_test_spoof, self.positives_test_spoof, eer_threshold_spoof)
            roc_points_dev_spoof     = bob.measure.roc(self.negatives_dev_spoof, self.positives_dev_spoof, 100)
            roc_points_test_spoof    = bob.measure.roc(self.negatives_test_spoof, self.positives_test_spoof, 100)

dev_hist_neg_spoof, dev_bin_neg_spoof = numpy.histogram(self.negatives_dev_spoof, self.n_bins_attack)
            test_hist_neg_spoof, test_bin_neg_spoof = numpy.histogram(self.negatives_test_spoof, self.n_bins_attack)
            
            # writes the output back to the platform
            output.write({
                'EER_licit': numpy.float32((far_dev_licit+frr_dev_licit) / 2.),
                'threshold_licit': numpy.float32(eer_threshold_licit),
                'FAR_dev_licit': numpy.float32(far_dev_licit),
                'FRR_dev_licit': numpy.float32(frr_dev_licit),
                'HTER_licit': numpy.float32((far_test_licit+frr_test_licit) / 2.),
                'FAR_test_licit': numpy.float32(far_test_licit),
                'FRR_test_licit': numpy.float32(frr_test_licit),
                'number_of_positives_dev_licit': numpy.int32(len(self.positives_dev_licit)),
                'number_of_negatives_dev_licit': numpy.int32(len(self.negatives_dev_licit)),
                'number_of_positives_test_licit': numpy.int32(len(self.positives_test_licit)),
                'number_of_negatives_test_licit': numpy.int32(len(self.negatives_test_licit)),
                'EER_spoof': numpy.float32((far_dev_spoof+frr_dev_spoof) / 2.),
                'threshold_spoof': numpy.float32(eer_threshold_spoof),
                'FAR_dev_spoof': numpy.float32(far_dev_spoof),
                'FRR_dev_spoof': numpy.float32(frr_dev_spoof),
                'HTER_spoof': numpy.float32((far_test_spoof+frr_test_spoof) / 2.),
                'FAR_test_spoof': numpy.float32(far_test_spoof),
                'FRR_test_spoof': numpy.float32(frr_test_spoof),
                'number_of_positives_dev_spoof': numpy.int32(len(self.positives_dev_spoof)),
                'number_of_negatives_dev_spoof': numpy.int32(len(self.negatives_dev_spoof)),
                'number_of_positives_test_spoof': numpy.int32(len(self.positives_test_spoof)),
                'number_of_negatives_test_spoof': numpy.int32(len(self.negatives_test_spoof)),
                'roc_dev': {
                    "data":
                        [
                            {
                                "label": "licit-protocol",
                                "false_positives": roc_points_dev_licit[0],
                                "false_negatives": roc_points_dev_licit[1],
                                "number_of_positives": numpy.uint64(len(self.positives_dev_licit)),
                                "number_of_negatives": numpy.uint64(len(self.negatives_dev_licit)),
                            },
                            {
                                "label": "spoof-protocol",
                                "false_positives": roc_points_dev_spoof[0],
                                "false_negatives": roc_points_dev_spoof[1],
                                "number_of_positives": numpy.uint64(len(self.positives_dev_spoof)),
                                "number_of_negatives": numpy.uint64(len(self.negatives_dev_spoof)),
                            }
                        ]
                    },
                'roc_test': {
                    "data":
                        [
                            {
                                "label": "licit-protocol",
                                "false_positives": roc_points_test_licit[0],
                                "false_negatives": roc_points_test_licit[1],
                                "number_of_positives": numpy.uint64(len(self.positives_test_licit)),
                                "number_of_negatives": numpy.uint64(len(self.negatives_test_licit)),
                            },
                            {
                                "label": "spoof-protocol",
                                "false_positives": roc_points_test_spoof[0],
                                "false_negatives": roc_points_test_spoof[1],
                                "number_of_positives": numpy.uint64(len(self.positives_test_spoof)),
                                "number_of_negatives": numpy.uint64(len(self.negatives_test_spoof)),
                            }
                        ]
                    },
                'dev_histograms': {
                    "data":
                        [
                            {
                                "label": "Attacks",
                                "x": dev_bin_neg_spoof[:-1],
                                "y": dev_hist_neg_spoof,
                            },
                            {
                                "label": "Zero-impostors",
                                "x": dev_bin_neg_licit[:-1],
                                "y": dev_hist_neg_licit,
                            },
                            {
                                "label": "Genuine",
                                "x": dev_bin_pos_licit[:-1],
                                "y": dev_hist_pos_licit,
                            }
                        ]
                    },
                'test_histograms': {
                    "data":
                        [
                            {
                                "label": "Attacks",
                                "x": test_bin_neg_spoof[:-1],
                                "y": test_hist_neg_spoof,
                            },
                            {
                                "label": "Zero-impostors",
                                "x": test_bin_neg_licit[:-1],
                                "y": test_hist_neg_licit,
                            },
                            {
                                "label": "Genuine",
                                "x": test_bin_pos_licit[:-1],
                                "y": test_hist_pos_licit,
                            }
                        ]
                    }
            })

return True

The code for this algorithm in Python
The ruler at 80 columns indicate suggested POSIX line breaks (for readability).
The editor will automatically enlarge to accomodate the entirety of your input
Use keyboard shortcuts for search/replace and faster editing. For example, use Ctrl-F (PC) or Cmd-F (Mac) to search through this box

An algorithm that implements standard metrics for biometric system evaluation.

Specifically, it returns:

eer: the equal error rate (EER) on the development set.
hter: the half total error rate (HTER) on the test set
far_dev: the false acceptance rate (FAR) on the development set
frr_dev: the false rejection rate (FRR) on the development set
far_dev: the false acceptance rate (FAR) on the test set
frr_test: the false rejection rate (FRR) on the test set
number_of_positives_dev: the number of positive (genuine) trials on the development set
number_of_negatives_dev: the number of negative (impostor) trials on the development set
number_of_positives_test: the number of positive (genuine) trials on the test set
number_of_negatives_test: the number of negative (impostor) trials on the test set
threshold: the threshold at the equal error rate on the development set
roc_dev: the receiver operating characteristic (ROC) curve on the development set according to the biometrics standard ISO/IEC 19795-1:2006(E)
roc_test: the receiver operating characteristic (ROC) curve on the test set according to the biometrics standard ISO/IEC 19795-1:2006(E)

This implementation relies on the &amp;#39;measure&amp;#39; package from the `Bob &amp;lt;http://www.idiap.ch/software/bob&amp;gt;`_ library.

Docutils System Messages

System Message: ERROR/3 (<string>, line 31); backlink

Unknown target name: "bob &amp;lt;http://www.idiap.ch/software/bob&amp;gt;".

Experiments

Attestation:

Privacy:

Status:

			Updated	Name	Databases/Protocols	Analyzers
				pkorshunov/pkorshunov/isv-asv-pad-fusion-complete/1/asv_isv-pad_lbp_hist_ratios_lr-fusion_lr-pa_aligned	avspoof/2@physicalaccess_verification,avspoof/2@physicalaccess_verification_spoof,avspoof/2@physicalaccess_antispoofing,avspoof/2@physicalaccess_verify_train_spoof,avspoof/2@physicalaccess_verify_train	pkorshunov/spoof-score-fusion-roc_hist/1
				pkorshunov/pkorshunov/isv-asv-pad-fusion-complete/1/asv_isv-pad_gmm-fusion_lr-pa	avspoof/2@physicalaccess_verification,avspoof/2@physicalaccess_verification_spoof,avspoof/2@physicalaccess_antispoofing,avspoof/2@physicalaccess_verify_train_spoof,avspoof/2@physicalaccess_verify_train	pkorshunov/spoof-score-fusion-roc_hist/1

Scientific Python 2.7 (0.1.0) 8

This table shows the number of times this algorithm has been successfully run using the given environment. Note this does not provide sufficient information to evaluate if the algorithm will run when submitted to different conditions.

algorithms pkorshunov spoof-score-fusion-roc_hist 1