Posted by
Herbie-3 on
Jan 22, 2014; 6:28pm
URL: http://imagej.273.s1.nabble.com/Jpeg-headers-not-saved-after-editing-the-image-tp5006218p5006235.html
Jon,
there are several aspects you didn't mention...
Yes there is lossless JPEG and there is no reason to criticize its use
but I'm not aware of someone really using it.
Compression strength in fact makes a difference but not if the task of
image processing or analysis is crucial. I've seen some false medical
decisions based on compression artifacts and I don't wish anybody to
suffer from a false diagnosis.
You are right, there are many, many other cases of image artifacts but
this doesn't justify to neglect those caused to JPEG compression.
Wrong:
[...] images with resolution far beyond what the physical resolution of
our machine justified. A little jpeg compression on top of this makes
no difference.
Physical resolution and jpeg compression artifacts are unrelated.
I stay with what I said and especially today there is no reason for
compressed images if they are used in scientific or critical context
such as a medical one. Storage is cheap and transmission bandwidth
became reasonable high even for large uncompressed images if they really
need to be transmitted in near to real-time.
A final aspect is that if you tolerate a certain and perhaps harmless
amount of JPEG artifacts you will be misunderstood by most of our fellow
scientist who are not image processing specialists. The result will be
that people think that JPEG compression poses no problem at all...
Finally, there is no reason to criticize JPEG compression with everyday
(consumer) camera use, surveillance purposes, etc.
BTW ImageJ opens and saves images with JPEG compression and there are
good reasons for this feature but such images are not suited for
scientific or otherwise critical purposes.
Best regards
Herbie
::::::::::::::::::::::::::::::::::::
On 22.01.14 18:53, Jon Harman wrote:
> Hi Herbie,
> I think that is too strong of a statement against jpegs. Whether jpeg
> is suitable or not depends on the analysis being done and the amount of
> compression being applied. The original imaging method is often
> responsible for gross distortions and the jpeg compression artifacts are
> minor in comparison.
>
> For example I used to work for a CT manufacturer. Radiologists
> routinely created images with resolution far beyond what the physical
> resolution of our machine justified. A little jpeg compression on top
> of this makes no difference.
>
> Jpeg has taken over imaging, for better or worse. (If only jpeg2000 had
> taken hold before the consumer digital cameras, alas.) ImageJ should
> support jpegs as well as it supports tiffs.
>
> Jon
>
>
>> BTW, be aware that JPEG-compressed images are unsuited for scientific
>> image processing and evaluation.
>>
>> HTH
>>
>> Herbie
>>
>
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html