I have absolutely no experience with any kind of AI and really want to create this:

A program that can train on a set of images to determine if an image is showing a fire flame or not (for fire detection). It should train on a set of "flame images".

I heard about the Keras python library which apparently allows to do this in 11 lines of code (http://adventuresinmachinelearning.com/keras-tutorial-cnn-11-lines/)...

Can anyone explain how I can make this work?

Thank you very much in advance!

  • $\begingroup$ Did you go through the tutorial and learn something? $\endgroup$ – Brian O'Donnell Apr 3 '18 at 1:12
  • $\begingroup$ Do you want somebody to read the article you posted to you? :P $\endgroup$ – solarflare Apr 4 '18 at 0:31
  • $\begingroup$ Oh so this article will completely answer my question? I can just use the entire code on there? $\endgroup$ – Vasya T Apr 4 '18 at 2:17
  • $\begingroup$ @VasyaT you will have to modify it, but yes. $\endgroup$ – FreezePhoenix Apr 10 '18 at 14:24

Many AI libraries allow the ability to feed them an Image, but others have to be modified to allow this.

Disregarding the information above, as it is just a notification, you can do this in many ways. Depending on the library you use, different methods give varying accuracy results.

The real questions are:

  1. How complex do you want the NN to be?
  2. Is this going to be commercial, or private?

For any of the answers above, you will have to find the best method that works for your needs. The more this matters to the world, the more complex it will have to be.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.