4 ms·
The example with the stop-sign on the building [0] really hits the nail on the head. For a human it is trivial to see that this is attached to a building (i.e.
by thecopy 5y ago
The example with the stop-sign on the building [0] really hits the nail on the head.
For a human it is trivial to see that this is attached to a building (i.e. not part of the road-sign network). I can imagine a camera-only system would struggle with this, and i wonder how a camera-only system can solve this in the general case without a breakthrough in general understanding and interpretation of the world.
[0] https://i0.wp.com/blog.piekniewski.info/wp-content/uploads/2021/05/E03PrXOXEAMOF38.jpeg?ssl=1 https://i0.wp.com/blog.piekniewski.info/wp-content/uploads/2...
- ben_w 5y agoIt’s trivial with any parallax, but in isolation and with only a glance I might have been fooled too. Also: https://youtu.be/r8j1JCVu6qY https://youtu.be/r8j1JCVu6qY
- thecopy 5y ago>It’s trivial with any parallax Sorry, but evidently it is not trivial, since Tesla failed it.
- ben_w 5y agoI meant for a human. If I’d seen that in the wild either while moving or with both eyes open, that wall sign wouldn’t have made me pause and think. Yet it was a single non-moving photo, so for a very brief moment it did.
- cupofcoffee 5y agoHumans, how do you build those?
- ben_w 5y agoTake two suitable Homo heidelbergensis…
- rasz 5y agoTesla uses monocular solution.