Integral Channel Features Detector

This section describes classes for object detection using WaldBoost and Integral Channel Features from [Šochman05] and [Dollár09].

computeChannels

Compute channels for integral channel features evaluation

C++: void computeChannels(InputArray image, vector<Mat>& channels)

Parameters:

• image – image for which channels should be computed
• channels – output array for computed channels

FeatureEvaluator

Feature evaluation interface

class FeatureEvaluator : public Algorithm

FeatureEvaluator::setChannels

Set channels for feature evaluation

C++: void FeatureEvaluator::setChannels(InputArrayOfArrays channels)

Parameters:

• channels – array of channels to be set

FeatureEvaluator::setPosition

Set window position to sample features with shift. By default position is (0, 0).

C++: void FeatureEvaluator::setPosition(Size position)

Parameters:

• position – position to be set

FeatureEvaluator::evaluate

Evaluate feature value with given index for current channels and window position.

C++: int FeatureEvaluator::evaluate(size_t feature_ind) const

Parameters:

• feature_ind – index of feature to be evaluated

FeatureEvaluator::evaluateAll

Evaluate all features for current channels and window position.

C++: void FeatureEvaluator::evaluateAll(OutputArray feature_values)

Parameters:

• feature_values – matrix-column of evaluated feature values

generateFeatures

Generate integral features. Returns vector of features.

C++: vector<vector<int>> generateFeatures(Size window_size, const string& type, int count=INT_MAX, int channel_count=10)

Parameters:

• window_size – size of window in which features should be evaluated
• type – feature type. Can be “icf” or “acf”
• count – number of features to generate.
• channel_count – number of feature channels

createFeatureEvaluator

Construct feature evaluator.

C++: Ptr<FeatureEvaluator> createFeatureEvaluator(const vector<vector<int>>& features, const string& type)

Parameters:

• features – features for evaluation
• type – feature type. Can be “icf” or “acf”

WaldBoostParams

Parameters for WaldBoost. weak_count — number of weak learners, alpha — cascade thresholding param.

struct CV_EXPORTS WaldBoostParams
{
    int weak_count;
    float alpha;

    WaldBoostParams(): weak_count(100), alpha(0.02f)
    {}
};

WaldBoost

class WaldBoost : public Algorithm

WaldBoost::train

Train WaldBoost cascade for given data. Returns feature indices chosen for cascade. Feature enumeration starts from 0.

C++: vector<int> WaldBoost::train(const Mat& data, const Mat& labels)

Parameters:

• data – matrix of feature values, size M x N, one feature per row
• labels – matrix of samples class labels, size 1 x N. Labels can be from {-1, +1}

WaldBoost::predict

Predict objects class given object that can compute object features. Returns unnormed confidence value — measure of confidence that object is from class +1.

C++: float WaldBoost::predict(const Ptr<FeatureEvaluator>& feature_evaluator) const

Parameters:

• feature_evaluator – object that can compute features by demand

WaldBoost::write

Write WaldBoost to FileStorage

C++: void WaldBoost::write(FileStorage& fs)

Parameters:

• fs – FileStorage for output

WaldBoost::read

Write WaldBoost to FileNode

C++: void WaldBoost::read(const FileNode& node)

Parameters:

• node – FileNode for reading

createWaldBoost

Construct WaldBoost object.

C++: Ptr<WaldBoost> createWaldBoost(const WaldBoostParams& params=WaldBoostParams())

ICFDetectorParams

Params for ICFDetector training.

struct CV_EXPORTS ICFDetectorParams
{
    int feature_count;
    int weak_count;
    int model_n_rows;
    int model_n_cols;
    int bg_per_image;
    std::string features_type;
    float alpha;
    bool is_grayscale;
    bool use_fast_log;

    ICFDetectorParams(): feature_count(UINT_MAX), weak_count(100),
        model_n_rows(56), model_n_cols(56), bg_per_image(5),
        alpha(0.02), is_grayscale(false), use_fast_log(false)
    {}
};

ICFDetector

class ICFDetector

ICFDetector::train

Train detector.

C++: void ICFDetector::train(const std::vector<String>& pos_filenames, const std::vector<String>& bg_filenames, ICFDetectorParams params=ICFDetectorParams())

Parameters:

• pos_path – path to folder with images of objects (wildcards like /my/path/*.png are allowed)
• bg_path – path to folder with background images
• params – parameters for detector training

ICFDetector::detect

Detect objects on image.

C++: void ICFDetector::detect(const Mat& image, vector<Rect>& objects, float scaleFactor, Size minSize, Size maxSize, float threshold, int slidingStep, std::vector<float>& values)

C++: detect(const Mat& img, std::vector<Rect>& objects, float minScaleFactor, float maxScaleFactor, float factorStep, float threshold, int slidingStep, std::vector<float>& values)

Parameters:

• image – image for detection
• objects – output array of bounding boxes
• scaleFactor – scale between layers in detection pyramid
• minSize – min size of objects in pixels
• maxSize – max size of objects in pixels
• minScaleFactor – min factor by which the image will be resized
• maxScaleFactor – max factor by which the image will be resized
• factorStep – scaling factor is incremented each pyramid layer according to this parameter
• slidingStep – sliding window step
• values – output vector with values of positive samples

ICFDetector::write

Write detector to FileStorage.

C++: void ICFDetector::write(FileStorage& fs) const

Parameters:

• fs – FileStorage for output

ICFDetector::read

Write ICFDetector to FileNode

C++: void ICFDetector::read(const FileNode& node)

Parameters:

• node – FileNode for reading

[Šochman05]

Šochman and J. Matas. WaldBoost – Learning for Time Constrained Sequential Detection”, CVPR, 2005. The paper is available online.

[Dollár09]

Dollár, Z. Tu, P. Perona and S. Belongie. “Integral Channel Features”, BMCV 2009. The paper is available online.