mbalatsko/opencv-rolling-ball

Fully Ported to Python from ImageJ's Background Subtractor. Rolling ball and sliding paraboloid background subtraction algorithms.

opencv-rolling-ball's Issues

Can not apply on the dicom images

The execution never stops when I pass a dicom image pixel array to its input. Can anyone tell the correct procedure to that?

My code:

import cv2
from cv2_rolling_ball import subtract_background_rolling_ball
import pydicom as dicom
...

dcm_img = dicom.dcmread("./left.dcm")

img, bg = subtract_background_rolling_ball(dcm_img.pixel_array, radius=50, light_background=False, use_paraboloid=True)

cv2.imshow('Original Image', img)
cv2.waitKey(0)

Return image and background at the same time?

After the background is computed, generating the final image is trivial.
I think, getting rid of create_background parameter and returning both outputs is more pythonic.
Snippet from example then will look like

img, background = subtract_background_rolling_ball(img, 30,
                                                   light_background=True,
                                                   use_paraboloid=False, 
                                                   do_presmooth=True)

Can it be faster?

Hello,

I like this package very much!

But its speed seems to be much more slower than using imageJ software. I wonder if there are any improvement space for speed?

Many thanks!