Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
falcor84
on May 12, 2024
|
parent
|
context
|
favorite
| on:
Whole-body magnetic resonance imaging at 0.05 Tesl...
To nitpick, wouldn't it by definition bias the read toward normal? I suppose the problem is more that you don't want to bias it to normal if it wasn't.
mminer237
on May 12, 2024
[–]
The training data is going to have far more normal scans of any given part than it will abnormal.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: