ITK Pre-Processing Algorithm Wrappers in Python/GrayscaleDilateFilter

Grayscale Dilate filter

This filter is implemented using the itkGrayscaleDilateImageFilter class in ITK. For each pixel, the filter finds the maximum of the pixels in a certain neighbourhood (defined by a structuring element). On application of this filter, some of the object expand in size and the image tends to look more bright than the original image. The default structuring element used is the Binary Ball structuring element (implemented using the itkBinaryBallStructuringElement class). The dilation is usually combined with erosion to form two powerful operations - morphological opening (erosion followed by dilation) and morphological closing (dilation followed by erosion).

Parameters:

• Radius - this specifies the radius of the structuring element to be used. Larger the value of the radius, larger the objects expand on dilation. Fig. 1 and Fig. 2 show how the objects expand and become brigter on dilation, with a radius of 1 and 3, respectively, when dilation is performed on seg_orig. Note that a lot of noise also seems to have emphasized in the image indicating that it is best not to dilate an image before doing some noise reduction.

Back to main page.

Fig. 1: Grayscale dilated image of seg_orig with radius 1

Fig. 2: Grayscale dilated image of seg_orig with radius 3