JPEG to DICOM

[Julius Klever]

I have some jpeg files that i would like to convert to dicom.

The images are conventional radiographs (14"x17") scanned with a "normal" transmitted light scanner.

File size is about 20MB (what would be ok i think) at a resolution of (i guess) 1000 dots per inch.

Unfortunatelly i don't know if any compression was used. But my Problem is:

If I use some GUI Tools to convert JPEG to DICOM (like the OsiriX Plugin) i'll get DICOM Files of several hundreds of MB.

If I use the dcm4che2 toolkit - DICOM Filesize is the same as JPEG, everything is fine, but now i have those funny moving stripes on the image (see attached video).

Does anybody know the reason? My idea is that the resolution of 1000dpi could be the problem? but i could not find any information according to the resolutions of dicom-images...

I must say I tested only with OsiriX, so could be an OsiriX specific problem too...

edit: seems to be the moiré pattern - anybody an idea how to improve that?

[Dimitri Pianeta]

Have-you use program as here http://samucs.blogspot.fr/2011/02/creating-dicom-multiframe-files-mpeg-to.html#more ?
For the size of image (jpeg), we can use method interpolator (say zoom in the image world). This method use to be polynomial (interpolator).
For instance, six degrees of spline interpolator or bicubic and so on.

  

[Rady]

Julius, 

If the Film size is 20 MB and it is RAW then 14X17 at 1000 dpi should be JPEG with good compression. I am not sure of you need such high dpi. Generally people do 200 to 300 dpi for scanning medical films. The reason why I wanted to point this is, many free convertors will fail to handle such volume of data. 

You have few that I can suggest for you to try. 

1. k-PACS (this is the best but 32bit, might be a problem for your file size)

2. Clearcanvas x64bit workstation with a Plugin. CC is now distributes only code. If you need I can send you. 

Hope this helps.

regards

Rady

[Julius Klever]

Thanks for your responses.
Im not a specialist in graphic - so first thing i will try is to reduce resolution (i dont know the x-ray film and intensifying screen that was used, but i guess the resolution of the conventional radiographs is not more than 10lp/mm what is about 500dpi)
so i think i will do some testscans at different resolutions.

but thanks for the other two tools - i didnt keep them in mind. (k-pacs is in fact very powerfull, but in my opinion a bit ugly :P , and clearcanvas was long time my favorite on windows, but its not the fastest viewer...)

best regards
julius

[zzzy...@gmail.com]

"I'm not a specialist in graphic - so first thing i will try is to reduce resolution."

As far as I know, image with a smaller image resolution will enhance the speed performance significantly, but with the trade off of the image quality.

For JPEG conversion to DICOM, you may check this article: http://samucs.blogspot.com/2008/12/converting-jpeg-to-dicom-using-dcm4che.html

在 2013年11月24日星期日UTC+8上午4时49分44秒，Julius Klever写道：

I have some jpeg files that i would like to convert to dicom.

The images are conventional radiographs (14"x17") scanned with a "normal" transmitted light scanner.

File size is about 20MB (what would be ok i think) at a resolution of (i guess) 1000 dots per inch.

Unfortunatelly i don't know if any compression was used. But my Problem is:

If I use some GUI Tools to convert JPEG to DICOM (like the OsiriX Plugin) i'll get DICOM Files of several hundreds of MB.

If I use the dcm4che2 toolkit - DICOM Filesize is the same as JPEG, everything is fine, but now i have those funny moving stripes on the image (see attached video).

Does anybody know the reason? My idea is that the resolution of 1000dpi could be the problem in image conversion? but i could not find any information according to the resolutions of dicom-images...