We’re having a challenge with finding any type of standards or guidelines in regards to altering medical images, whether using lossless or visually lossless, and have decided to factor those out completely (we believe only a very small percentage of the PDF files have embedded images – approx. 2 – 3 %).
Once we factor the PDF files with images, then we feel that we can compress the remainder of the files using lossy compression, to get the largest amount of compression, without risk.
We’re still trying to sort through this, so by all means, please let us know if we’re overlooking anything.
Thanks.