Classifier.h

#ifndef CLASSIFIER_HEADER
#define CLASSIFIER_HEADER

// #include "FeatureExtractor.h"
#include "IntegralImage.h"

class Classifier {
public:
    Classifier();
    virtual ~Classifier();

    virtual int Classify(const IntegralImage *intImage, const Rect &roi, float scale = 1.0f) {
        return -1;
    }

    virtual float Evaluate(const IntegralImage *intImage, const Rect &roi, float scale = 1.0f) {
        return FLT_MIN;
    }
};

/*
// Define the structure of features.
// Notice that the only the last weak classifier
// in a stage use the threshold field.
// Keep it in the structure is for memory alignment.
#define H_OFFSET        0
#define W_OFFSET        1
#define HEIGHT_OFFSET   2
#define WIDTH_OFFSET    3
#define Z_OFFSET        4
#define MIN_OFFSET      5
#define MAX_OFFSET      6
#define THRE_OFFSET     7
#define PROJ_OFFSET     8
#define HISTO_OFFSET    44
#define NUM_PROJ        36
#define NUM_HISTO       100

// Total size of a weak classifier.
#define SIZE_WEAK       144

/*************************************
The structure is like this:

struct WeakClassifier {
feat h;
feat w;
feat height;
feat width;
feat z;
feat min;
feat max;
feat threshold;
feat projection[36];
feat histogram[100];
};

****************************************************/

/*
class AdaBoostClassifier : public Classifier {
public:
    // Build the AdaBoost Classifier from files.
    AdaBoostClassifier(const char *filepath);

    ~AdaBoostClassifier();

    // Classify the patch.
    // @param x: the x coordinate of upperleft corner.
    // @param y: the y coordinate of upperleft corner.
    // @param featureExt: the Feature Extractor.
    bool Classify(int x, int y, feat scale, FeatureExtractor &featureExt);

private:
    // Data.
    int numStages;
    int *numWeaks;
    feat *data;

    inline int District(feat value, feat min, feat max) const {
        if (value <= min) return 0;
        if (value >= max) return 99;

        feat step = (max - min) * 0.01;

        for (int i = 0; i < 100; i++)
            if (value >= min + step * i - 0.000001 && value <= min + step * (i + 1) + 0.000001)
                return i;       
    }

    // A Feature instance to store the feature.
    Feature feature;
};

*/
#endif