The only constructor

obj = cv.StructuredEdgeDetection(model)
obj = cv.StructuredEdgeDetection(model, howToGetFeatures)

Input

• model name of the file where the model is stored.
• howToGetFeatures optional name of MATLAB M-function that implements custom feature extractor. A helper function for training part of: "P. Dollar and C. L. Zitnick. Structured Forests for Fast Edge Detection, 2013". You need it only if you would like to train your own forest, otherwise leave it unspecified for the default implementation. See example below.

Example

The following is an example of a custom feature extractor MATLAB function:

% This function extracts feature channels from src. The
% StructureEdgeDetection uses this feature space to detect
% edges.
function features = myRFFeatureGetter(src, opts)
    % src: source image to extract features
    % features: output n-channel floating-point feature matrix
    % opts: struct of options
    gnrmRad = opts.gradientNormalizationRadius;
    gsmthRad = opts.gradientSmoothingRadius;
    shrink = opts.shrinkNumber;
    outNum = opts.numberOfOutputChannels;
    gradNum = opts.numberOfGradientOrientations;

    nsize = [size(src,1) size(src,2)] ./ shrink;
    features = zeros([nsize outNum], 'single');
    % ... here your feature extraction code
end

TODO: Custom extractor is not internally used in the current cv.StructuredEdgeDetection implementation. See http://docs.opencv.org/3.1.0/d2/d59/tutorial_ximgproc_training.html for more information about training your own structured forest (it uses an external MATLAB toolbox for the training part).