Is CAPCHA used to train Full self driving vision?

Is CAPCHA used to train Full self driving vision?

Ever noticed how CAPCHA has lots of questions surrounding transportation.

- Lights
- Intersection
- Hills
- Cars
- Bikes
- Buses
- Bridges
- Store fronts
- Traffic signs

Sounds like stuff you would want your full self driving car to be able to distinguish.

reed_lewis | February 11, 2019

I doubt it. If that were the use, then if you selected the wrong pictures it would not let you continue on the web page. Imagine if the first person to see a new picture selected the wrong elements, then forever that picture would never work.

No these pictures are ones that have already been marked with attributes (stop lights, store fronts, hills, etc.) and then they present you a number of pictures and some of them have the item.

El Mirio | February 11, 2019

Did some digging, it seems they do use these type of human data inputs to train AI.

On your point tho, maybe the picture is shown simultaneously to many human users at the same time, so likelihood of wrong tag is reduced by majority.

reed_lewis | February 22, 2019

Wow, my response has been used in a SPAM post. I feel honored! I flagged it.