Detection mask

A compound tool that splits an image to fixed-size blocks and produces a low-resolution detection mask image by classifying each sub-image. The sub-image classifier is provided by an inherited implementation.

Inputs

• image: The image in which objects are to be detected.

• gridSpacing: The horizontal and vertical spacing between grid corner points in pixel coordinates. A 1-by-2 matrix.

• gridPosition: Specifies how the grid should be positioned. See Generate grid.

Outputs

• mask: A low-resolution mask image whose pixel intensities are classifications/regressions/probabilities produced by the internal classifier, scaled to [0, 255].

enum GridPosition

Different ways of positioning the grid.

Values:

enumerator FitInImage

Make sure each cell fits in the image completely.

If the size of the image is not a multiple of gridSpacing, leave equal top/bottom and left/right margins. If the input image is smaller than gridSpacing, this technique produces an empty set of points.

enumerator CoverWholeImage

Make sure every pixel in the image is in one of the cells.

If the size of the image is not a multiple of gridSpacing, the top/bottom and left/right cells exceed image boundaries by an equal amount.