how to do right pair prediction in python from left image and depthmap ,it will be very helpful , if you find in solution python ,gpu.
Thanks..

Got in working on Linux Mint 19.2, by changing in .local/lib/python2.7/site-packages/PIL/Image.py:

Line 1968: rotn_center = (int(w / 2.0), int(h / 2.0))
Line 2387: self.im.transform2(( int(box[0]),int(box[1]),int(box[2]),int(box[3]) ), image.im, method, data, resample, fill)
Line 2495: return Image()._new(core.new(mode, (int(size[0]),int(size[1])) ))

  File "/Users/cmahnke/projects/StereoscoPy/stereoscopy/__init__.py", line 370, in resize
    image = image.resize(re_size, Image.ANTIALIAS)
AttributeError: module 'PIL.Image' has no attribute 'ANTIALIAS'

This has changed with Pillow 10, see https://pillow.readthedocs.io/en/stable/releasenotes/10.0.0.html#constants

Please add left.jpg and right.jpg to the example_images directory to be able to try out the provided examples from README.md directly.

Calling this:

stereoscopy.create_anaglyph((left, right), color_scheme="red-green", method="gray").save(rgAnagFileName)

Results in the following:

Shouldn't the background be white / gray, like calling without the color_scheme param? That is just green where the are cyan parts without the color_scheme param.

After searching for example it seems that the orange tint is actually to be expected...
Can anyone confirm?

I like this project. I used it to make a puzzle for a treasure hunt for my wife. I see that there has not been any activity for a while, so this may fall on deaf ears. However, I was working in an environment where OpenCV (4.2.0 I believe) was already installed. I pip installed StereoscoPy on a Windows machine. When I first tried to run the script, there was an error in the call to findTransformECC. It wanted an inputMask and a gaussian blur radius. I modified the code to pass None and 1 respectively, and it worked.