1

Is there a way to apply gaussian blurring to an image where each pixel is blurred using its own standard deviation which is passed as a mask? I want to be able to apply stronger blurring to certain parts of the image depending on some condition. Ideally, this would be something like blurred=variable_gauss(image (HxWx3), std_mask (HxW))

Dennis Sakva
  • 1,447
  • 2
  • 13
  • 26
  • You could simply use a sliding window which computes each "condition" as it passes and applies the blur. – JP1 Feb 10 '17 at 09:41
  • That would be painfully slow in python. By condition I just mean a pre-calculated mask of floats which shows the strength of blurring. Actually, I don't really need blurring to be gaussian. It's just that gaussian blurring looks good. – Dennis Sakva Feb 10 '17 at 09:49
  • What will be the SD of the blur? only to be applied to the voxel? in that case you could simply apply a matrix multiplication if i understand correctly – JP1 Feb 10 '17 at 10:05
  • The mask would contain SD for individual pixels. The x3 refers to RGB channels (not real voxels and gaussian does not need to be 3D), so blurring would be applied equally to each channel. It's not really needed there. – Dennis Sakva Feb 10 '17 at 10:23
  • Possible duplicate of [How to blur some portion of Image in Android?](http://stackoverflow.com/questions/30101044/how-to-blur-some-portion-of-image-in-android) – Spektre Feb 10 '17 at 12:01

0 Answers0