Detects objects of different sizes in the input image

boxes = classifier.detect(im)
[boxes, numDetections] = classifier.detect(im)
[boxes, rejectLevels, levelWeights] = classifier.detect(im)
[...] = classifier.detect(im, 'OptionName', optionValue, ...)

Input

• im Matrix of the type uint8 containing an image where objects are detected.

Output

• boxes Cell array of rectangles where each rectangle contains the detected object, the rectangles may be partially outside the original image.
• numDetections optional vector of detection numbers for the corresponding objects. An object's number of detections is the number of neighboring positively classified rectangles that were joined together to form the object.
• rejectLevels optional output vector of integers. Implies OutputRejectLevels=true.
• levelWeights optional output vector of doubles. Implies OutputRejectLevels=true.

Options

• ScaleFactor Parameter specifying how much the image size is reduced at each image scale. default 1.1
• MinNeighbors Parameter specifying how many neighbors each candiate rectangle should have to retain it. default 3
• MinSize Minimum possible object size. Objects smaller than that are ignored. Not set by default.
• MaxSize Maximum possible object size. Objects larger than that are ignored. If MaxSize == MinSize model is evaluated on single scale. Not set by default.
• OutputRejectLevels if is true returns rejectLevels and levelWeights. default false
• DoCannyPruning Parameter with the same meaning for an old cascade as in the function cvHaarDetectObjects. It is not used for a new cascade. default false
• ScaleImage Parameter with the same meaning for an old cascade as in the function cvHaarDetectObjects. It is not used for a new cascade. default false
• FindBiggestObject Parameter with the same meaning for an old cascade as in the function cvHaarDetectObjects. It is not used for a new cascade. default false
• DoRoughSearch Parameter with the same meaning for an old cascade as in the function cvHaarDetectObjects. It is not used for a new cascade. default false

The detected objects are returned as a cell array of rectangles. Note that the function has three variants based on the number of output arguments.

The function is parallelized with the TBB library.

The third variant allows you to retrieve the final stage decision certainty of classification. For this, one needs to set OutputRejectLevels to true and request the rejectLevels and levelWeights output arguments. For each resulting detection, levelWeights will then contain the certainty of classification at the final stage. This value can then be used to separate strong from weaker classifications. A code sample on how to use it efficiently can be found below:

model = cv.CascadeClassifier('/path/to/your/model.xml');
[boxes, levels, weights] = model.detect(img, 'OutputRejectLevels',true);
fprintf('Detection [%d,%d,%d,%d] with weight %f\n', boxes{1}, weights(1));