Ready for groupRectangles refactor

Scott Klum
1 parent 08179e6d
Showing 4 changed files with 22 additions and 21 deletions
openbr/openbr_plugin.h
openbr/plugins/classification/boostedforest.cpp
openbr/plugins/classification/cascade.cpp
openbr/plugins/imgproc/slidingwindow.cpp
@@ -1422,7 +1422,7 @@ public:
     static Classifier *make(QString str, QObject *parent); /*!< \brief Make a classifier from a string. */
     virtual void train(const QList<cv::Mat> &images, const QList<float> &labels) = 0;
-    virtual float classify(const cv::Mat &image, float *confidence = NULL) const = 0;
+    virtual float classify(const cv::Mat &image, bool process = true, float *confidence = NULL) const = 0;
     // Slots for representation
     virtual cv::Mat preprocess(const cv::Mat &image) const = 0;
@@ -129,18 +129,24 @@ class BoostedForestClassifier : public Classifier
         }
     }
-    float classify(const Mat &image, float *confidence) const
+    float classify(const Mat &image, bool process, float *confidence) const
     {
+        Mat m;
+        if (process)
+            m = preprocess(image);
+        else
+            m = image;
+
         float sum = 0;
         for (int i = 0; i < classifiers.size(); i++) {
             Node *node = classifiers[i];
             while (node->left) {
                 if (representation->maxCatCount() > 1) {
-                    int c = (int)representation->evaluate(image, node->featureIdx);
+                    int c = (int)representation->evaluate(m, node->featureIdx);
                     node = (node->subset[c >> 5] & (1 << (c & 31))) ? node->left : node->right;
                 } else {
-                    double val = representation->evaluate(image, node->featureIdx);
+                    double val = representation->evaluate(m, node->featureIdx);
                     node = val <= node->threshold ? node->left : node->right;
                 }
             }
@@ -152,11 +152,11 @@ class CascadeClassifier : public Classifier
         }
     }
-    float classify(const Mat &image, float *confidence) const
+    float classify(const Mat &image, bool process, float *confidence) const
     {
         float stageConf = 0.0f;
         foreach (const Classifier *stage, stages) {
-            float result = stage->classify(image, &stageConf);
+            float result = stage->classify(image, process, &stageConf);
             if (result == 0.0f) {
                 if (confidence)
                     *confidence += stageConf;
@@ -219,7 +219,7 @@ private:
             if (!imgHandler.getPos(pos))
                 qFatal("Cannot get another positive sample!");
-            if (classify(pos, &confidence) > 0.0f) {
+            if (classify(pos, true, &confidence) > 0.0f) {
                 printf("POS current samples: %d\r", images.size());
                 images.append(pos);
                 labels.append(1.0f);
@@ -235,7 +235,7 @@ private:
             if (!imgHandler.getNeg(neg))
                 qFatal("Cannot get another negative sample!");
-            if (classify(neg, &confidence) > 0.0f) {
+            if (classify(neg, true, &confidence) > 0.0f) {
                 printf("NEG current samples: %d\r", images.size() - posCount);
                 images.append(neg);
                 labels.append(0.0f);
@@ -53,12 +53,6 @@ class SlidingWindowTransform : public MetaTransform
     BR_PROPERTY(QString, model, "")
-    void init()
-    {
-        QDataStream stream;
-        load(stream);
-    }
-
     void train(const TemplateList &data)
     {
         classifier->train(data.data(), File::get<float>(data, "Label", -1));
@@ -90,8 +84,7 @@ class SlidingWindowTransform : public MetaTransform
                 Mat m;
                 OpenCVUtils::cvtUChar(t[i], m);
                 std::vector<Rect> rects;
-                std::vector<int> rejectLevels;
-                std::vector<double> levelWeights;
+                std::vector<float> confidences;
                 if (maxObjectSize.height == 0 || maxObjectSize.width == 0)
                     maxObjectSize = m.size();
@@ -123,21 +116,23 @@ class SlidingWindowTransform : public MetaTransform
                         for (int x = 0; x < processingRectSize.width; x += step) {
                             Mat window = repImage(Rect(Point(x, y), Size(originalWindowSize.width + dx, originalWindowSize.height + dy))).clone();
-                            float gypWeight;
-                            int result = classifier->classify(window, &gypWeight);
+                            float confidence;
+                            int result = classifier->classify(window, false, &confidence);
-                            if (12 - result < 4) {
+                            if (result == 1) {
                                 rects.push_back(Rect(cvRound(x*factor), cvRound(y*factor), windowSize.width, windowSize.height));
-                                rejectLevels.push_back(result);
-                                levelWeights.push_back(gypWeight);
+                                confidences.push_back(confidence);
                             }
+                            // Add ROC mode
+
                             if (result == 0)
                                 x += step;
                         }
                     }
                 }
+                // groupRectangles(rects, confidences, eps);
                 groupRectangles(rects, rejectLevels, levelWeights, minNeighbors, eps);
                 if (!enrollAll && rects.empty())