CaptionBot trys to identify Linus

[Doodlegame]

If you havent heard of captionbot- Microsofts new AI ( I think they ditched Tay for a while ) scans your pictures and provides a caption on what the picture is based on facial recognition body tracking like on xbox kinnect (Yay for he spelling Microsoft). Anyway I was mucking about on this new service from Microsoft adn it was great! It could even recognise a Llama! 

However, when I uploaded the attached picture that is below it just couldnt work it out... I wondor why

[Doodlegame]

Just now, Starelementpoke said:

Probably the text.

Thats what I think but I thought it was still funny as it is meant to read text now haha. I have just written a mini revew for it in the review forum; go and check it out if you are interested

[samcool55]

Lol i uploaded a picture of 3 anime cat faces and it responded with: "I am not really confident, but I think it's a person on a surf board in a skate park."

Who the hell would take a surf board to a skate park? O_o

 

This thing is hilarious, it did manage to do a good job with a confused travolta.

Let's mess with it a bit more.

[Kinda Bottlenecked]

Quote

I am not really confident, but I think it's person wearing a hat.

 

Helmet. Okay, pretty close. 

[Doodlegame]

9 minutes ago, samcool55 said:

Lol i uploaded a picture of 3 anime cat faces and it responded with: "I am not really confident, but I think it's a person on a surf board in a skate park."

Who the hell would take a surf board to a skate park? O_o

 

This thing is hilarious, it did manage to do a good job with a confused travolta.

Let's mess with it a bit more.

Haha ill try some more stuff with it

[Starelementpoke]

12 minutes ago, Doodlegame said:

Thats what I think but I thought it was still funny as it is meant to read text now haha. I have just written a mini revew for it in the review forum; go and check it out if you are interested

Well, to be fair, all the text in the photo is either an abbreviation or misspelled.

[Doodlegame]

Just now, Starelementpoke said:

Well, to be fair, all the text in the photo is either an abbreviation or misspelled.

I suppose so good point

[samcool55]

8 minutes ago, Pohernori said:

 

Helmet. Okay, pretty close. 

That sounds not too bad, i've uploaded a map and somehow it said it was a map next to some giraffes, no idea where he sees giraffes but oh well

It's interesting how it sees certain things in very basic photo's.

 

edit: i uploaded a picture of a box, and somehow it recognized donuts in a box next to a brick wall.

Just upload the most simple shapes and you confuse the crap out of it

[Dredgy]

It's surprisingly accurate. Not perfect, but for a computer it's definitely picking up  most aspects of a scene.

[Doodlegame]

Well, Microsoft made the Xbox 360 AND CaptionBot; Of course they are compatable hahahahah

[Doodlegame]

1 minute ago, samcool55 said:

That sounds not too bad, i've uploaded a map and somehow it said it was a map next to some giraffes, no idea where he sees giraffes but oh well

It's interesting how it sees certain things in very basic photo's.

 

edit: i uploaded a picture of a box, and somehow it recognized donuts in a box next to a brick wall.

Just upload the most simple shapes and you confuse the crap out of it

Haha i get the same things with it

[Coaxialgamer]

Got a few ...interesting results . But hey , at least it recognized my picture of bill cosby

 

 

Here is some random stuff i got with pictures of linus

Quote

Quote

 

[Doodlegame]

38 minutes ago, Coaxialgamer said:

Got a few ...interesting results . But hey , at least it recognized my picture of bill cosby

 

 

Here is some random stuff i got with pictures of linus