JpgXr compression for fluorescent images

Post your acquisition-related questions and macros here
Post Reply
user-6053
Posts: 9
Joined: Thu Jan 01, 1970 1:00 am

JpgXr compression for fluorescent images

Post by user-6053 »

Dear Zeiss software specialist,

I'm using ZEN2.3 slidescan to acquire some multi-channel fluorescent images of my fixed tissue sample slides. And before I used another Zeiss microscope on similar samples but different treatment groups. I aim to compare the result from them, and treat the images acquired from the two Zeiss systems in the same way.

Lately I found the images acquired from the slide scanner are with a "JpgXr" compression and 85 compression quality (see attached screeshots). I am wondering if that going to affect my downstream image analysis and maybe even the scientific conclusion. Particularly I'm interested in the local variance (or standard deviation of fluorescent signal) within a cell, and need to quantify standard deviation of the different fluorescent channels of all cells and compare them among groups.
I have heard jpeg compression is lossy and not good for scientific research, so it's a bit hard to guess why there is such mode provided by Zeiss on a research instrument. And I couldn't find too much documentation about the JpgXr format from Zeiss.

Could you share a little more information about this format, and suggest if it's fine to use it for my project? If it should be avoided at all time, it's also important for me to know.

Thanks.
Ziqiang Huang
Attachments
compressed.jpg
compressed.jpg (93.36 KiB) Viewed 1128 times
uncompressed.jpg
uncompressed.jpg (97.73 KiB) Viewed 1128 times
user-4
Posts: 398
Joined: Thu Jan 01, 1970 1:00 am

Post by user-4 »

Dear Ziqiang Huang,

We did a lot of testing with various compression methods. At the end we came to the conclusion that "JpgXr" compression offers us the best ratio between minimal loss of image quality and maximum reduction of file size in the case of ‘little values’ for the selected compression.

With 85 compression quality we get reduced filesize by a factor of 10 to 20. In case of RGB this is ‘not visible’ and for fureszence images it has the ‘positive effect’ of noise reduction.

But this might be different for your special case.

I think the best way for you is to acquire an image without compression, then save the image again with compression and finally compare these images.
Post Reply