Re: Image Newbie Question
- From: "Roger Hamlett" <rogerspamignored@xxxxxxxxxxxxxxxxxxx>
- Date: Wed, 07 Dec 2005 22:11:33 GMT
"Musashi" <musashi@xxxxxxxxxx> wrote in message
news:%gIlf.6$Ax.2@xxxxxxxxxxxxx
>I have a JPG of some books in a bookcase
> (from a b&w picture taken in the 50's)
> and I'd like to be able to recover the
> names on the books.
>
> When I zoom in on the jpg, the letters are too blurred
> to read.
>
> Could someone point me to the correct image
> processing algorithm to try and recover the
> titles on the books in this jpg.
>
> Would it be some sort of low pass filter.
> I am a software person, so once the
> correct algorithm is suggested,
> I can implement it.
>
> Thanks
> M
Realistically, you probably can't. It depends how close you are to being
able to see the image anyway. These 'Holywood' systems, where images from
a low definition CCTV camera, are miraculously sharpened to reveal
thousands of times the original picture quality, are pure fiction. Now
there will be two seperate problems. The original optical blurring from
the imaging system, and the jpeg artefacts. If you can get hold of the
original file, not using jpeg, you will have a much better chance. Some of
the possibilities, will depend on what JPEG format was used. Standard
JPEG, handles the data in fixed size 'blocks', which leads to pixelation
of the resulting image. This loss is permanent. Hwever variants like
JPEG2000, attempt to reduce this problem. If you can get access to the
original data (this means before JPEG compression, converting the JPEG
image to another better format, will not restore the data that has been
lost), then a deconvolution algorithm, or a high pass filter may well show
edges that are otherwise not visible. Jpeg Image enhancer, will remove the
visible blockiness, but the data loss is still present. _If_ there is just
enough resolution, it is possible that this, followed by a deconvolution,
might improve things,but you really have to reckon that the data is not
there.
Best Wishes
.
- References:
- Image Newbie Question
- From: Musashi
- Image Newbie Question
- Prev by Date: Image Newbie Question
- Next by Date: Re: Image Newbie Question
- Previous by thread: Image Newbie Question
- Next by thread: Re: Image Newbie Question
- Index(es):
Relevant Pages
|