Sign in to follow this  

Low Pass image filter

This topic is 4837 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Take the image, do the low-pass (gaussian blur), subtract the blurred version from the original version. Bingo: high pass.

Share this post


Link to post
Share on other sites
Thanks for that advise. Is that the fastest way?

oh and what do you mean by subtract? EG if the original version has a black pixel and the corresponding pixel on the blurred version is dark gray?

Share this post


Link to post
Share on other sites
If you choose to use the blur-subtract method, be sure to do so in signed number space, and normalize the result to 0...1 space after the operation, since the result may well be negative. Photoshop's high pass filter, for example, normalizes so that 50% gray is the "empty" band color by adding [0.5, 0.5, 0.5] (but this could be controlled very easily with another bias factor during normalizing).

Laplacian matrix is indeed equivalent to the blur-subtract method described here, and is actually faster since the core operation needs only one step (and the bias as needed).

Share this post


Link to post
Share on other sites
Quote:
Original post by Nodger
Ok, this "LaPlacian Mask" sounds like the way to go... Can anyone suggest any good URLs with tuts for this?

it's convoluting with kernel

0 1 0
1 -4 1
0 1 0

Or similar kernel(maybe with some values instead of 0's, maybe multiplied by some constant, etc). Basically a Gaussian blur kernel but with center equal to -(summ of other cells).

Note that if you're working on pixels storen in one byte per color component, you might need to add 127 , you'll have negative results.

http://homepages.inf.ed.ac.uk/rbf/HIPR2/log.htm

[google] search for Laplacian kernel

Share this post


Link to post
Share on other sites

This topic is 4837 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this