Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

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.


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: