Sign in to follow this  

image artifact removal for cartoons

This topic is 2043 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

Is there any information about techniques for removing compression artifacts from cartoon like images? What little information I've been able to find about compression artifact removal focuses on photos. However, it seems like it should be possible to do much better with cartoon like images due to the knowledge that the image mostly consists of large flat areas and sharp lines.

Share this post


Link to post
Share on other sites
I don't know if this is the right section. Anyway. Is you problem specific to compressed textures or just to images? Because certain image formats have loss-less compression (PNG for example). With compressed textures I don't know if you can solve those artifacts. About tecniques I think there's a lot of material on the web about image compression tecniques/ algorithms.

Probably Vectorial Art can interest you since is a total different way for store images and have no artifacts indipendently of resolution at wich you display the image.

EDIT: probably I misunderstood you.

Removing artifacts can be done with GIMP (the first thing come to my mind is to do some gaussian blur with small radius, and then use sharpenizer). Probably someone has also written a plugin for that too.

This goes in 2d and visual art. Edited by DemonRad

Share this post


Link to post
Share on other sites
This is math, so perhaps "Math and Physics" would have been more appropriate.

I've seen some very spectacular demonstrations of a technique called "basis pursuit" that could probably be used for this purpose. I don't know of anything specific to removing compression artifacts from cartoon images, but I happen to work with two of the people that invented the technique, so I'll ask them to see if they know of any relevant papers.

Share this post


Link to post
Share on other sites
If you know the number of colors that are used in the cartoon image then k-means clustering the colors with k being the number colors usually works pretty well. If you want it to be fully-automated or if you want the output to be anti-aliased then you need to do something more sophisticated; for example, anisotropic diffusion filtering (google it).

This problem is closely related to the general problem of raster-to-vector conversion in that it is basically the first step of raster-to-vector conversion i.e. image segmentation, so there's a lot of literature on the subject. Google image segmentation and/or raster-to-vector. Edited by jwezorek

Share this post


Link to post
Share on other sites
Another option would be:

Build a library of artifacts based and their original uncompressed 8x8 value.

JPEG, and MPEG I think, only use 8x8 blocks when compressing.

I've seen this trick used before to deblurr masked digits (like when someone posts a picture of a cheque and blurs the account number). You could probably get away with using greyscale and then determine the colors once a matching artifact is found.

There is likely enough structural information present for you to reverse most of the compression.. Since its a cartoon. :)

Share this post


Link to post
Share on other sites

This topic is 2043 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