Whether you're a fan of skiing, hiking, snowshoeing, or just a lover of sunbathing in the restaurant on the slopes, I'm sure you've already wondered what mountain was in front of you.
With this new application, you will be asked a simple photo to answer this question. You can then show off your new knowledge in front of your date, your friends or your family !
For the most curious, it will even be possible to have certain characteristics of the mountain identified : its height, its dangerousness, how to get there, whether it is possible to ski, walk or not...
For example, could you give me the names of those famous mountains before using the application?
To do that, first we are going to use
ml5.js,
"a JS library that aims to make machine learning approachable for a broad audience of artists, creative coders, and students." In addition of ml5.js, we use Machine Learning and MobileNets, a network which permits to recognize images : an essential tool for our website. By using these two modules, IA can recognize a picture and give a percentage of confidence.
For example, with a bird :
This recognition is also used by a lot of devices, as Apple with its facial recognition to unlock a phone.
To create a website like this, you can use this code below. This is the code of the same example above with the bird.
All these information are in a javascript file. The first thing to do, it is to define variables.
The first one is the image of the mountain you want to analyse or classify. The second one is called "classifier" and represents the MobileNet. Then, you define functions, to predict what mountain it is and the probability of the result.
We have seen that IA can recognize a mountain, but it has to be taught sometimes.
So to improve this, don't hesitate to give it new pictures of mountains you know and the names associated !