rand-convolve 1.0.5

Runs a multi-random-kernel convolution on an image.

This package provides the ability to run a random convolution on a given image.

Main functions:

• rand_convolve:

  • Returns a randomly convoluted image given an image path as input.

  • Arguments:

    • image: The image path for the convolution.
    • ksize: The size of the 2 dimensional kernel. The shape of the kernel is: (size, size). (default: 3, must be an odd number.)
    • knum: The number of random kernels to generate and use. (default: 20)
      • int n>0: The number of kernels.
      • -1: Use a new kernel for each pixel.
    • ksign: The signs to be used for the kernels. (default: 0)
      • 0: All kernel values are positive.
      • 1: Kernel signs are randomly mixed between negative and positive.
      • 2: All kernel values are negative.

  • Returns:

    • A PIL.Image object of the convoluted image.