let classifier;

​

​

  // Model URL (!!replace with your model URL!!) 

  let imageModelURL = 'https://teachablemachine.withgoogle.com/models/1fK5h8jX/';

​

  // Video

  let video;

  let flippedVideo;

  // To store the classification

  let label = "";

  let port;

  let connect_button;

​

  // Load the model first

  function preload() {

    classifier = ml5.imageClassifier(imageModelURL + 'model.json');

  }

​

  function setup() {

    createCanvas(600, 600);

    // Create the video

    video = createCapture(VIDEO);

    video.size(320, 240);

    video.hide();

    connect_button = createButton('connect arduino');

    connect_button.position(width/2, height / 1.5);

    connect_button.id('connect');

​

    flippedVideo = ml5.flipImage(video)

    // Start classifying

    classifyVideo();

  }

​

  function draw() {

    background(0);

    // Draw the video

    image(flippedVideo, 0, 0);

​

    // Draw the label

    fill(255);

    textSize(16);

    textAlign(CENTER);

    text(label, width / 2, height - 4);

  }

​

  // Get a prediction for the current video frame

  function classifyVideo() {

    flippedVideo = ml5.flipImage(video)

    classifier.classify(flippedVideo, gotResult);

  }

​

  // When we get a result

  function gotResult(error, results) {

    // If there is an error

    if (error) {

      console.error(error);

      return;

    }

    // The results are in an array ordered by confidence.

    // console.log(results[0]);

    label = results[0].label;

    data_to_send = new Uint8Array(1);

    if (label === 'My Class Label 1') {

      data_to_send[0] = 1;  

    }

    else if (label === 'My Class Label 2') {

      data_to_send[0] = 2;  

    }

    else if (label === 'My Class Label 3') {

      data_to_send[0] = 3;  

    }

    if (port != undefined) {

      port.send(data_to_send);

    }

    // Classifiy again!

    classifyVideo();

  }