Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
FINN: A Framework for Fast, Scalable Binarized Neural Network Inference [pdf] (ntnu.no)
7 points by lucodibidil on Nov 29, 2016 | hide | past | favorite | 1 comment


They report image classification speeds of 20K - 12.3 million images / second. Good! That means we might see robots that don't stumble or take 2 minutes to open a door. Here's what you can do with 1000fps vision:

https://www.youtube.com/watch?v=-KxjVlaLBmk

(and that was 7 years ago!)

I'm wondering why fast vision applied to robotics hasn't been more researched in the meantime.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: