Image Bandpass Filter

Bandpass filter for 2D images. Remove high- and low-frequency noise and stripes. Butterworth, gaussian, and hard filter options.

imbandpass

Bandpass filter for images (2D arrays) supressing high-frequency noise, low-frequency variations, and stripes. Gaussian, Butterworth, or hard filter options.

To use: call function imbandpass(image, low_cutoff, high_cutoff), returns smoothed image.

The default options of Gaussian filter, Gaussian stripe filter, and mirrored padding were chosen to replicate imageJ's FFT Bandpass filter.

To access non-default options including stripe supression, use keyword arguments, for example

imbandpass(I, 3, 250, filter="Butterworth", stripes="Horizontal", stripeTolerance=10)

equivalently

imbandpass(I, 3, 250, "filter", "Butterworth", "stripes", "Horizontal", "stripeTolerance", 10).

Arguments

Positional

image Image in. Handles single-channel or RGB images as arrays: input (m,n) or (m,n,3) array of values in range 0 to 255.

low_cutoff - filter out features below this (real space) lengthscale in pixels.

high_cutoff - filter out features above this lengthscale.

It's possible to set upper and/or lower cutoff to [] and not apply this aspect of the filter.

Optional keyword parameters

stripes = 'Horizontal' ,'Vertical', or 'None' - supress stripes, default 'None'

stripeFilter = 'Gaussian' or 'hard' - stripe filter mode, default 'gaussian'

stripeTolerance - tolerance (in percent) for stripe deviation from horizontal/vertical alignement, default 5.

filter='gaussian', 'butterworth' or 'hard', filter profile, default 'gaussian'

butterworthN exponent in butterworth filter, default 1

padOption = 'symmetric' , 'replicate', 0 or other value, or 'None' - how to pad image border for Fourier transform, default 'symmetric'

Output

image_out - a uint8 array with the same dimensions and number of channels as image.

---

If this is useful to you or you want it fixed, please let me know with a rating/comment that it worked/how it didn't work!