11 messages in net.java.dev.jai-imageio.interestRe: [JAI-IMAGEIO] Problem with 12-bit...| From | Sent On | Attachments |
|---|---|---|
| Marco Sambin - NeoLogica | Nov 22, 2007 6:56 am | |
| robert engels | Nov 22, 2007 7:56 am | |
| Marco Sambin - NeoLogica | Nov 22, 2007 8:33 am |  .jpg |
| James Cheng | Nov 22, 2007 1:32 pm | |
| Marco Sambin - NeoLogica | Nov 23, 2007 2:33 am | |
| James Cheng | Dec 3, 2007 7:53 pm | |
| Marco Sambin - NeoLogica | Dec 4, 2007 2:04 am | |
| James Cheng | Dec 4, 2007 9:59 am | |
| jai-...@javadesktop.org | Dec 6, 2007 6:13 pm | |
| Marco Sambin - NeoLogica | Dec 7, 2007 1:19 am | |
| jai-...@javadesktop.org | Dec 11, 2007 2:32 pm |
| Subject: | Re: [JAI-IMAGEIO] Problem with 12-bit grayscale lossless JPEG image | |
|---|---|---|
| From: | robert engels (reng...@ix.netcom.com) | |
| Date: | Nov 22, 2007 7:56:36 am | |
| List: | net.java.dev.jai-imageio.interest | |
Nothing I have, including libjpeg can read the image.
They all error with:
Unsupported JPEG process: SOF type 0xc3
On Nov 22, 2007, at 8:56 AM, Marco Sambin - NeoLogica wrote:
Hi.
I am experiencing some problems when trying to decode a lossless JPEG grayscale image through the native JPEG decoder contained in the JAI Image Image I/O Tools. This grayscale lossless JPEG image is extracted from a DICOM medical image file.
The decoded image shall contain pixel values ranging from 0 (min pixel value) to 713 (max pixel value), according to the meta-information contained in the original DICOM dataset, and also according to what other image viewers display when requested to load the same sample image.
But, when decoded by the JIIO lossless JPEG decoder (through JAI's "ImageRead" operator), the resulting image contains pixel values ranging from 0 to 65533, which seems to be incorrect. The final result after contrast adjustment is an "all-white" image, and not the correct MR medical image, as displayed by other viewers.
The sample lossless JPEG image can be downloaded from:
http://test.neologica.it/Lossless.jpg
Do you believe what I noticed may derive from a decoding bug of the lossless JPEG JIIO Image Reader?
I have also prepared a self-contained test case which loads the image and displays it after the contrast adjustment suggested in the original DICOM dataset, and which is incorrectly displayed as all-white. Please let me know if you are interested in this test case.
Best regards,
Marco Sambin.
--------------------------------------------------------------------- To unsubscribe, e-mail: inte...@jai-imageio.dev.java.net For additional commands, e-mail: interest-help@jai- imageio.dev.java.net