Evaluate specified ROI and return confidence value for each location in multiple scales

[rcts, locations] = hog.detectMultiScaleROI(im, locations)
[...] = hog.detectMultiScaleROI(..., 'OptionName',optionValue, ...)

Input

• im 8-bit 1- or 3-channel image where objects are detected.
• locations input detection region of interest. It specifies candidate locations to search for object detections at different scales. An struct array with the following fields:
  • scale scale (size) of the bounding box, scalar.
  • locations set of requrested locations to be evaluated, cell array of points {[x,y], ...}.
  • confidences vector that will contain confidence values for each location. Not required on input, this will be filled/updated on output.

Output

• rcts Detected objects boundaries. Cell array of rectangles where objects are found, of the form {[x,y,w,h], ...}.
• locations output updated locations struct array. All points are retained, but their confidences are updated.

Options

• HitThreshold Threshold for the distance between features and SVM classifying plane. Usually it is 0 and should be specfied in the detector coefficients (as the last free coefficient). But if the free coefficient is omitted (which is allowed), you can specify it manually here. default 0
• GroupThreshold Minimum possible number of rectangles in a group minus 1. The threshold is used on a group of rectangles to decide whether to retain it or not. If less than or equal to zero, no grouping is performed. See cv.groupRectangles. default 0