Merge pull request #7 from biometrics/kde

Implemented Imposter Score Normalization and Match Probability

Merge pull request #7 from biometrics/kde
Implemented Imposter Score Normalization and Match Probability
jklontz
2 parents f74da8dd f9c9e2d1
Showing 6 changed files with 232 additions and 58 deletions
sdk/core/core.cpp
sdk/openbr_plugin.cpp
sdk/openbr_plugin.h
sdk/plugins/algorithms.cpp
sdk/plugins/compare.cpp
sdk/plugins/quality.cpp
@@ -252,7 +252,7 @@ private:
         if (words.size() > 2) qFatal("AlgorithmCore::init invalid algorithm format.");
         transform = QSharedPointer<Transform>(Transform::make(words[0], NULL));
-        if (words.size() > 1) distance = QSharedPointer<Distance>(Factory<Distance>::make("." + words[1]));
+        if (words.size() > 1) distance = QSharedPointer<Distance>(Distance::make(words[1], NULL));
     }
 };
@@ -497,6 +497,8 @@ QString Object::argument(int index) const
         return "[" + strings.join(",") + "]";
     } else if (type == "br::Transform*") {
         return variant.value<Transform*>()->description();
+    } else if (type == "br::Distance*") {
+        return variant.value<Distance*>()->description();
     } else if (type == "QStringList") {
         return "[" + variant.toStringList().join(",") + "]";
     }
@@ -524,6 +526,8 @@ void Object::store(QDataStream &amp;stream) const
                 transform->store(stream);
         } else if (type == "br::Transform*") {
             property.read(this).value<Transform*>()->store(stream);
+        } else if (type == "br::Distance*") {
+            property.read(this).value<Distance*>()->store(stream);
         } else if (type == "bool") {
             stream << property.read(this).toBool();
         } else if (type == "int") {
@@ -556,6 +560,8 @@ void Object::load(QDataStream &amp;stream)
                 transform->load(stream);
         } else if (type == "br::Transform*") {
             property.read(this).value<Transform*>()->load(stream);
+        } else if (type == "br::Distance*") {
+            property.read(this).value<Distance*>()->load(stream);
         } else if (type == "bool") {
             bool value;
             stream >> value;
@@ -620,6 +626,8 @@ void Object::setProperty(const QString &amp;name, const QString &amp;value)
         }
     } else if (type == "br::Transform*") {
         variant.setValue(Transform::make(value, this));
+    } else if (type == "br::Distance*") {
+        variant.setValue(Distance::make(value, this));
     } else if (type == "QStringList") {
         variant.setValue(parse(value.mid(1, value.size()-2)));
     } else if (type == "bool") {
@@ -630,6 +638,7 @@ void Object::setProperty(const QString &amp;name, const QString &amp;value)
     } else {
         variant = value;
     }
+
     if (!QObject::setProperty(qPrintable(name), variant) && !type.isEmpty())
         qFatal("Failed to set %s::%s to: %s %s",
                 metaObject()->className(), qPrintable(name), qPrintable(value), qPrintable(type));
@@ -786,6 +795,7 @@ void br::Context::initializeQt(QString sdkPath)
     qRegisterMetaType< QList<int> >();
     qRegisterMetaType< br::Transform* >();
     qRegisterMetaType< QList<br::Transform*> >();
+    qRegisterMetaType< br::Distance* >();
     qRegisterMetaType< cv::Mat >();
     qInstallMsgHandler(messageHandler);
@@ -1270,42 +1280,17 @@ void Transform::backProject(const TemplateList &amp;dst, TemplateList &amp;src) const
 }
 /* Distance - public methods */
-void Distance::train(const TemplateList &templates)
-{
-    const TemplateList samples = templates.mid(0, 2000);
-    const QList<float> sampleLabels = samples.labels<float>();
-    QScopedPointer<MatrixOutput> memoryOutput(dynamic_cast<MatrixOutput*>(Output::make(".Matrix", FileList(samples.size()), FileList(samples.size()))));
-    compare(samples, samples, memoryOutput.data());
-
-    double genuineAccumulator, impostorAccumulator;
-    int genuineCount, impostorCount;
-    genuineAccumulator = impostorAccumulator = genuineCount = impostorCount = 0;
-
-    for (int i=0; i<samples.size(); i++) {
-        for (int j=0; j<i; j++) {
-            const float val = memoryOutput.data()->data.at<float>(i, j);
-            if (sampleLabels[i] == sampleLabels[j]) {
-                genuineAccumulator += val;
-                genuineCount++;
-            } else {
-                impostorAccumulator += val;
-                impostorCount++;
-            }
-        }
-    }
-
-    if (genuineCount == 0) { qWarning("No genuine matches."); return; }
-    if (impostorCount == 0) { qWarning("No impostor matches."); return; }
-
-    double genuineMean = genuineAccumulator / genuineCount;
-    double impostorMean = impostorAccumulator / impostorCount;
-
-    if (genuineMean == impostorMean) { qWarning("Genuines and impostors are indistinguishable."); return; }
+Distance *Distance::make(QString str, QObject *parent)
+{
+        // Check for custom transforms
+        if (Globals->abbreviations.contains(str))
+            return make(Globals->abbreviations[str], parent);
-    a = 1.0/(genuineMean-impostorMean);
-    b = impostorMean;
+        File f = "." + str;
+        Distance *distance = Factory<Distance>::make(f);
-    qDebug("a = %f, b = %f", a, b);
+        distance->setParent(parent);
+        return distance;
 }
 void Distance::compare(const TemplateList &target, const TemplateList &query, Output *output) const
@@ -1336,7 +1321,7 @@ float Distance::compare(const Template &amp;target, const Template &amp;query) const
                 return -std::numeric_limits<float>::max();
         }
-    return a * (_compare(target, query) - b);
+    return _compare(target, query);
 }
 QList<float> Distance::compare(const TemplateList &targets, const Template &query) const
@@ -971,7 +971,6 @@ public:
 protected:
     Transform(bool independent = true); /*!< \brief Construct a transform. */
     inline Transform *make(const QString &description) { return make(description, this); } /*!< \brief Make a subtransform. */
-
 };
 /*!
@@ -1065,19 +1064,19 @@ class BR_EXPORT Distance : public Object
 {
     Q_OBJECT
-    // Score normalization
-    Q_PROPERTY(float a READ get_a WRITE set_a RESET reset_a)
-    Q_PROPERTY(float b READ get_b WRITE set_b RESET reset_b)
-    BR_PROPERTY(float, a, 1)
-    BR_PROPERTY(float, b, 0)
-
 public:
+    virtual ~Distance() {}
+    static Distance *make(QString str, QObject *parent); /*!< \brief Make a distance from a string. */
+
     static QSharedPointer<Distance> fromAlgorithm(const QString &algorithm); /*!< \brief Retrieve an algorithm's distance. */
-    virtual void train(const TemplateList &src); /*!< \brief Train the distance. */
+    virtual void train(const TemplateList &src) { (void) src; } /*!< \brief Train the distance. */
     virtual void compare(const TemplateList &target, const TemplateList &query, Output *output) const; /*!< \brief Compare two template lists. */
     float compare(const Template &target, const Template &query) const; /*!< \brief Compute the normalized distance between two templates. */
     QList<float> compare(const TemplateList &targets, const Template &query) const; /*!< \brief Compute the normalized distance between a template and a template list. */
+protected:
+    inline Distance *make(const QString &description) { return make(description, this); } /*!< \brief Make a subdistance. */
+
 private:
     virtual void compareBlock(const TemplateList &target, const TemplateList &query, Output *output, int targetOffset, int queryOffset) const;
     virtual float _compare(const Template &a, const Template &b) const = 0; /*!< \brief Compute the distance between two templates. */
@@ -1132,6 +1131,7 @@ Q_DECLARE_METATYPE(QList&lt;float&gt;)
 Q_DECLARE_METATYPE(QList<int>)
 Q_DECLARE_METATYPE(br::Transform*)
 Q_DECLARE_METATYPE(QList<br::Transform*>)
+Q_DECLARE_METATYPE(br::Distance*)
 Q_DECLARE_METATYPE(cv::Mat)
 #endif // __OPENBR_PLUGIN_H
@@ -70,6 +70,7 @@ class AlgorithmsInitializer : public Initializer
         Globals->abbreviations.insert("FaceClassificationExtraction", "((Grid(7,7)+SIFTDescriptor(8)+ByRow)/DenseLBP+PCA(0.95,instances=-1)+Cat)");
         Globals->abbreviations.insert("AgeRegressor", "Center(Range,instances=-1)+SVM(RBF,EPS_SVR,instances=100)");
         Globals->abbreviations.insert("GenderClassifier", "Center(Range,instances=-1)+SVM(RBF,C_SVC,instances=4000)");
+        Globals->abbreviations.insert("UCharL1", "Unit(ByteL1)");
     }
 };
@@ -119,7 +119,7 @@ BR_REGISTER(Distance, DistDistance)
  * \brief Fast 8-bit L1 distance
  * \author Josh Klontz \cite jklontz
  */
-class UCharL1Distance : public Distance
+class ByteL1Distance : public Distance
 {
     Q_OBJECT
@@ -129,7 +129,7 @@ class UCharL1Distance : public Distance
     }
 };
-BR_REGISTER(Distance, UCharL1Distance)
+BR_REGISTER(Distance, ByteL1Distance)
 /*!
@@ -137,7 +137,7 @@ BR_REGISTER(Distance, UCharL1Distance)
  * \brief Fast 4-bit L1 distance
  * \author Josh Klontz \cite jklontz
  */
-class PackedUCharL1Distance : public Distance
+class HalfByteL1Distance : public Distance
 {
     Q_OBJECT
@@ -147,7 +147,7 @@ class PackedUCharL1Distance : public Distance
     }
 };
-BR_REGISTER(Distance, PackedUCharL1Distance)
+BR_REGISTER(Distance, HalfByteL1Distance)
 /*!
  * \ingroup distances
@@ -18,14 +18,20 @@ class IUMTransform : public Transform
     Q_PROPERTY(br::Distance* distance READ get_distance WRITE set_distance RESET reset_distance STORED false)
     Q_PROPERTY(double mean READ get_mean WRITE set_mean RESET reset_mean)
     Q_PROPERTY(double stddev READ get_stddev WRITE set_stddev RESET reset_stddev)
-    BR_PROPERTY(br::Distance*, distance, Factory<Distance>::make(".Dist(L2)"))
+    BR_PROPERTY(br::Distance*, distance, Distance::make("Dist(L2)", this))
     BR_PROPERTY(double, mean, 0)
     BR_PROPERTY(double, stddev, 1)
     br::TemplateList impostors;
     float calculateIUM(const Template &probe, const TemplateList &gallery) const
     {
-        QList<float> scores = distance->compare(gallery, probe);
+        const int probeLabel = probe.file.label();
+        TemplateList subset = gallery;
+        for (int j=subset.size()-1; j>=0; j--)
+            if (subset[j].file.label() == probeLabel)
+                subset.removeAt(j);
+
+        QList<float> scores = distance->compare(subset, probe);
         float min, max;
         Common::MinMax(scores, &min, &max);
         double mean;
@@ -39,14 +45,8 @@ class IUMTransform : public Transform
         impostors = data;
         QList<float> iums; iums.reserve(impostors.size());
-        QList<int> labels = impostors.labels<int>();
-        for (int i=0; i<data.size(); i++) {
-            TemplateList subset = impostors;
-            for (int j=subset.size()-1; j>=0; j--)
-                if (labels[j] == labels[i])
-                    subset.removeAt(j);
-            iums.append(calculateIUM(impostors[i], subset));
-        }
+        for (int i=0; i<data.size(); i++)
+            iums.append(calculateIUM(impostors[i], impostors));
         Common::MeanStdDev(iums, &mean, &stddev);
     }
@@ -74,6 +74,194 @@ class IUMTransform : public Transform
 BR_REGISTER(Transform, IUMTransform)
+/* Kernel Density Estimator */
+struct KDE
+{
+    float min, max;
+    double mean, stddev;
+    QList<float> bins;
+
+    KDE() : min(0), max(1) {}
+    KDE(const QList<float> &scores)
+    {
+        Common::MinMax(scores, &min, &max);
+        Common::MeanStdDev(scores, &mean, &stddev);
+        double h = Common::KernelDensityBandwidth(scores);
+        const int size = 255;
+        bins.reserve(size);
+        for (int i=0; i<size; i++)
+            bins.append(Common::KernelDensityEstimation(scores, min + (max-min)*i/(size-1), h));
+    }
+
+    float operator()(float score, bool gaussian = false) const
+    {
+        if (gaussian) return 1/(stddev*sqrt(2*CV_PI))*exp(-0.5*pow((score-mean)/stddev, 2));
+        if (score <= min) return bins.first();
+        if (score >= max) return bins.last();
+        const float x = (score-min)/(max-min)*bins.size();
+        const float y1 = bins[floor(x)];
+        const float y2 = bins[ceil(x)];
+        return y1 + (y2-y1)*(x-floor(x));
+    }
+};
+
+QDataStream &operator<<(QDataStream &stream, const KDE &kde)
+{
+    return stream << kde.min << kde.max << kde.mean << kde.stddev << kde.bins;
+}
+
+QDataStream &operator>>(QDataStream &stream, KDE &kde)
+{
+    return stream >> kde.min >> kde.max >> kde.mean >> kde.stddev >> kde.bins;
+}
+
+/* Match Probability */
+struct MP
+{
+    KDE genuine, impostor;
+    MP() {}
+    MP(const QList<float> &genuineScores, const QList<float> &impostorScores)
+        : genuine(genuineScores), impostor(impostorScores) {}
+    float operator()(float score, bool gaussian = false, bool log = false) const
+    {
+        const float g = genuine(score, gaussian);
+        const float s = g / (impostor(score, gaussian) + g);
+        if (log) return (std::max(std::log10(s), -10.f) + 10)/10;
+        else     return s;
+    }
+};
+
+QDataStream &operator<<(QDataStream &stream, const MP &nmp)
+{
+    return stream << nmp.genuine << nmp.impostor;
+}
+
+QDataStream &operator>>(QDataStream &stream, MP &nmp)
+{
+    return stream >> nmp.genuine >> nmp.impostor;
+}
+
+/*!
+ * \ingroup distances
+ * \brief Match Probability \cite klare12
+ * \author Josh Klontz \cite jklontz
+ */
+class MPDistance : public Distance
+{
+    Q_OBJECT
+    Q_PROPERTY(br::Distance* distance READ get_distance WRITE set_distance RESET reset_distance STORED false)
+    Q_PROPERTY(QString binKey READ get_binKey WRITE set_binKey RESET reset_binKey STORED false)
+    Q_PROPERTY(bool gaussian READ get_gaussian WRITE set_gaussian RESET reset_gaussian STORED false)
+    Q_PROPERTY(bool log READ get_log WRITE set_log RESET reset_log STORED false)
+    BR_PROPERTY(br::Distance*, distance, make("Dist(L2)"))
+    BR_PROPERTY(QString, binKey, "")
+    BR_PROPERTY(bool, gaussian, false)
+    BR_PROPERTY(bool, log, false)
+
+    QHash<QString, MP> mps;
+
+    void train(const TemplateList &src)
+    {
+        distance->train(src);
+
+        const QList<int> labels = src.labels<int>();
+        QScopedPointer<MatrixOutput> memoryOutput(dynamic_cast<MatrixOutput*>(Output::make(".Matrix", FileList(src.size()), FileList(src.size()))));
+        distance->compare(src, src, memoryOutput.data());
+
+        QHash< QString, QList<float> > genuineScores, impostorScores;
+        for (int i=0; i<src.size(); i++)
+            for (int j=0; j<i; j++) {
+                const float score = memoryOutput.data()->data.at<float>(i, j);
+                const QString bin = src[i].file.getString(binKey, "");
+                if (labels[i] == labels[j]) genuineScores[bin].append(score);
+                else                        impostorScores[bin].append(score);
+            }
+
+        foreach (const QString &key, genuineScores.keys())
+            mps.insert(key, MP(genuineScores[key], impostorScores[key]));
+    }
+
+    float _compare(const Template &target, const Template &query) const
+    {
+        return mps[query.file.getString(binKey, "")](distance->compare(target, query), gaussian, log);
+    }
+
+    void store(QDataStream &stream) const
+    {
+        distance->store(stream);
+        stream << mps;
+    }
+
+    void load(QDataStream &stream)
+    {
+        distance->load(stream);
+        stream >> mps;
+    }
+};
+
+BR_REGISTER(Distance, MPDistance)
+
+/*!
+ * \ingroup distances
+ * \brief Linear normalizes of a distance so the mean impostor score is 0 and the mean genuine score is 1.
+ * \author Josh Klontz \cite jklontz
+ */
+class UnitDistance : public Distance
+{
+    Q_OBJECT
+    Q_PROPERTY(br::Distance *distance READ get_distance WRITE set_distance RESET reset_distance)
+    Q_PROPERTY(float a READ get_a WRITE set_a RESET reset_a)
+    Q_PROPERTY(float b READ get_b WRITE set_b RESET reset_b)
+    BR_PROPERTY(br::Distance*, distance, make("Dist(L2)"))
+    BR_PROPERTY(float, a, 1)
+    BR_PROPERTY(float, b, 0)
+
+    void train(const TemplateList &templates)
+    {
+        const TemplateList samples = templates.mid(0, 2000);
+        const QList<float> sampleLabels = samples.labels<float>();
+        QScopedPointer<MatrixOutput> memoryOutput(dynamic_cast<MatrixOutput*>(Output::make(".Matrix", FileList(samples.size()), FileList(samples.size()))));
+        compare(samples, samples, memoryOutput.data());
+
+        double genuineAccumulator, impostorAccumulator;
+        int genuineCount, impostorCount;
+        genuineAccumulator = impostorAccumulator = genuineCount = impostorCount = 0;
+
+        for (int i=0; i<samples.size(); i++) {
+            for (int j=0; j<i; j++) {
+                const float val = memoryOutput.data()->data.at<float>(i, j);
+                if (sampleLabels[i] == sampleLabels[j]) {
+                    genuineAccumulator += val;
+                    genuineCount++;
+                } else {
+                    impostorAccumulator += val;
+                    impostorCount++;
+                }
+            }
+        }
+
+        if (genuineCount == 0) { qWarning("No genuine matches."); return; }
+        if (impostorCount == 0) { qWarning("No impostor matches."); return; }
+
+        double genuineMean = genuineAccumulator / genuineCount;
+        double impostorMean = impostorAccumulator / impostorCount;
+
+        if (genuineMean == impostorMean) { qWarning("Genuines and impostors are indistinguishable."); return; }
+
+        a = 1.0/(genuineMean-impostorMean);
+        b = impostorMean;
+
+        qDebug("a = %f, b = %f", a, b);
+    }
+
+    float _compare(const Template &target, const Template &query) const
+    {
+        return a * (distance->compare(target, query) - b);
+    }
+};
+
+BR_REGISTER(Distance, UnitDistance)
+
 } // namespace br
 #include "quality.moc"