Class ObjectDetector

Initialize an ObjectDetector.

Input:

object_type: the type of object to detect,

in ('frontal_face','profile_face','upper_body','lower_body','full_body')

flag: the type of operation of the object detector

0 = use opencv's default haar cascade 1 = use the cascade stored in a BinaryClassifier, see param1, param2 2 = same as 1 but use opencv's default haar cascade to post-verify 3 = use the cascade stored in a text file named by param1 4 = same as 3 but use opencv's default haar cascade to post-verify

param1: [optional] a strong classifier to classify a patch into object/non-object

If flag is 1 or 2: This holds a BinaryClassifier If flag is 3 or 4: This holds a path to the text file storing the cascade

param2: [optional] If flag is 1 or 2:

this holds the length of the image patch (e.g. 20-by-20 or 24-by-24)

block_size : int

length of a square block. An image is divided into blocks if it is too large. This offers fast and stable object detection.

Detect objects from a grayscale image.

Input:
    img: a POINTER(IplImage) of depth IPL_DEPTH_8U and 1 or 3 channels
    scale_factor: The factor by which the search window is scaled 
        between the subsequent scans, for example, 1.1 means increasing window by 10%.
    min_neighbors: Minimum number (minus 1) of neighbor rectangles that makes up an object. 
        All the groups of a smaller number of rectangles than min_neighbors-1 are rejected. 
        If min_neighbors is 0, the function does not any grouping at all and returns all 
        the detected candidate rectangles, which may be useful if the user wants to apply 
        a customized grouping procedure.
    group_overlapping: if True then overlapping locations are grouped
Output:
    z: an 'int' numpy.array of tuples (x,y,w,h) representing the object locations