In case others are not aware what "pathology scanner" is, apparently it is a device to scan/image microscope slides. Found some specs, apparently these Philips units do 0.25um/px and 15mm x 15mm imaging area, making the output images presumably 60000 x 60000 pixels in size. Apparently Philips previously used their own "iSyntax" format, and also JPEG2000 DICOM files for these devices.
https://connect.mozilla.org/t5/ideas/support-jpeg-xl/idi-p/1...
vote for this feature to be natively supported in browsers
Seeing JPEG XL integrated into the DICOM standard was a particularly proud moment for me as the manager of the effort at Google.
It felt like closing a major circle in my career, because I spent the first 16 years of my career in the medical industry, working on Neurosurgical Robots (Oulu Neuronavigator System), and one of the first tools I built was an ACR-NEMA 1.0 parser, ACR-NEMA being the direct predecessor to DICOM, and then continuing on radiation treatment planning systems with plenty of DICOM work within them. To now contribute back to that very standard is incredibly rewarding.
I find Philips quite an interesting player in imaging, PACS and related space. At the innovation space they might be coming up with new technologies and solutions but during service and delivery side they provide quite an awful service, at least in my personal experience in Singapore. I have friends at various healthcare institutions and we were just surprised how could the service be so bad for a vendor who is considered as a valuable one in the industry.
My first ever job in software was working for PathXL (a Belfast startup implementing digital pathology software). Lots of fond memories working there, including how cool it was working on what was effectively Google Maps but for massive tissue sample images. PathXL actually ended up getting acquired by Philips, seems like a great match if they're building the hardware for this.
Can someone comment on what is newsworthy about this?
After reading a bit about jpeg xl [0], the bit depth, channel count and pixel count seem promising. Devils in the details. How will multiple focal planes be implemented?
https://www.abyssmedia.com/heic-converter/avif-heic-jpegxl.s...
Always impressed when someone does anything with DICOM, it's a bit complex format IMHO.
WebP artifacts not pathological enough?
Ugh, Pathology image processing is really annoying.
IF Philips is going to stick to the DICOM format, and not add lots of proprietary stuff, _and_ it's the format that it uses internally, then this will be good.
For example, folks can check out OpenSlide (https://openslide.org) and have a look at all the different slide formats that exist. If you dig in to Philips' entry, you'll see that OpenSlide does not support Philips' non-TIFF format (iSyntax), and that the TIFF format is an "export format".
If you have a Philips microscope that uses iSyntax, you are very limited on what non-Philips software you can use. If you want files in TIFF format, you (the lab tech) have to take an action to export a side in TIFF. It can take up a fair amount of lab tech time.
Ideally, the microscope should immediately store the images in an open format, with metadata that workflow software can use to check if a scanning run is complete. I _hope_ that will be able to happen here!