Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
p13rr0m
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
4 ms
·
1.
▲
by
p13rr0m
7y ago
I believe the paper SlowFast Networks for Video Recognition goes somewhat in that direction. The architecture is split into a fast pathway, to capture motion information, and a slow pathway that captures spatial semantics. https:/