6 ms·
I remember a researcher doing some early research on compressing diagnostic imaging and was happy about all the hard disk space saved. They did some research to
by Scoundreller 6y ago
I remember a researcher doing some early research on compressing diagnostic imaging and was happy about all the hard disk space saved. They did some research to find out what level of compression they could go with that wouldn't result in different clinicians reaching different conclusions from the same images.
It really upset me. We probably threw away decades of training data that a computer could have used for early detection.
Fine for broken arms of whatever, but for cancer diagnostics, ugh. The computer might have been able to see the tumour before a clinician.
- SubiculumCode 6y agoIt is easy to forget, but disk space used to be a super precious resource for labs.
- grogenaut 6y agoThat data compression meant getting imagery to experts in less than an hour digitally instead of via sneakernet physically in a multi-day timeframe, thus massively speeding up and improving access to high quality diagnosis.
- Scoundreller 6y agoCould do both, but still depends on storage costs. But I guess re-scanning wouldn't get justified once there's already a "good enough" compressed version. Real question is if they dialed down the compression as storage got cheaper. Doubtful.