How to make a neural network on images?

The challenge is to make a neural network which is able to recognize the game in its image.

That is, at the input of the data we have with the image of the game (screenshot) and therefore what the game is(game Name).

I want to train a neural network to it in the future, could the screenshot to define the game.
Accordingly I want to get a tip what to read (what a neural network is suitable, why and how to implement it).
Actually what language is it implemented?What tools to use and what existing solutions there are? Myself am a java developer.
June 10th 19 at 14:44
1 answer
June 10th 19 at 14:46
I want to get a tip


Borman gave Stirlitz tip
"A little, but in foreign currency" - thought Stirlitz


the problem is
Actually what language to implement it

to explain so that you understand

brief - take any network search by images
they are usually the same in Python (well, with the engine compartment siplyusplyus)
Python is also necessary as-that to explain what to look for. Can only experienced gamers something can determine. - alfonso56 commented on June 10th 19 at 14:49
Oh
a set of screenshots + how Yandex is looking for films

strictly speaking, it is without neuronti can be done, but it is stronger - Juanita17 commented on June 10th 19 at 14:52
In General, I imagine that what, from the point of view of the theory(I read a lot about Alpago), but interested in practical implementation. Tips, examples, on the basis of which we can do on our own.

How do you imagine to do without nareki? Of interest is that let's say a new game, and starts to appear a new array of the data online (on the fly) and he is trained.

That is, to achieve the performance necessary underhood C++? - orie.Russ commented on June 10th 19 at 14:55

Find more questions by tags Neural networks