First, the attention grabbing bit - using traditional digital image blurring techniques may not be effective in all cases. The title of the article from which this post derives is "Why blurring sensitive information is a bad idea".
Why? With statistical techniques, it may be possible to extract some of the original image details were. Now for the complicating details. First, it wasn't shown with a real case - it was simulated. Second, although one could probably deduce this from the article, traditional digital blurring is in all likelihood very effective against information with high spatial frequencies (i.e. images with lots of changes, like pictures of faces not taken at a close distance). The problem MIGHT occur with blurring data with mostly low frequency content, like large sized numbers, such as on an image of the check used in the article.
Nonetheless, it is a very interesting piece of work. I look forward to someone showing this work against "real data".
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment