//KNN variables

//------

console.log("ml5 version:", ml5.version);

​

let isModelReady = false;

let isVideoReady = false;

let isClassifying = false;

​

// let video;

const knnClassifier = ml5.KNNClassifier();

let featureExtractor;

​

let resultLabel = "";

let resultScore = 0.0;

​

let label0_score = 0.0;

let label1_score = 0.0;

let label2_score = 0.0;

let label0_count = 0;

let label1_count = 0;

let label2_count = 0;

​

let isAMirrorSelfie = {

  label_score: 0.0,

  label_count: 0,

  label_text: "This is a mirror selfie.",

};

​

let notAMirrorSelfie = {

  label_score: 0.0,

  label_count: 0,

  label_text: "This is NOT a mirror selfie.",

};

​

//--------------------------------------------------------------

// KNN functions

//--------------------------------------------------------------

​

function displayModelInfo(){  

  let yShift = 60;

  fill(0, 255, 0);

  push();

  textSize(20);

  if (resultLabel === ""){

    text("I'm not guessing anything until you press the pink button" , 10, 20 + yShift);

  } else {

    text("Guessing it is a " + resultLabel + ": " + resultScore, 10, 20 + yShift);

  }

  pop();

​

  text(

    "Mirror Selfie:  " + label0_count + "  |  " + floor(label0_score) + "%",

    10,

    50 + yShift

  );

  text(

    "Not Mirror Selfie:  " + label1_count + "  |  " + floor(label1_score) + "%",

    10,

    70 + yShift

  );

  text(

    "Label 2:  " + label2_count + "  |  " + floor(label2_score) + "%",

    10,

    90 + yShift

  );

}

​

function modelReady() {

  console.log("Model Loaded: FeatureExtractor");

  isModelReady = true;

  loadKNNDataset();

}

​

function videoReady() {

  console.log("Device Ready");

  isVideoReady = true;

  loadKNNDataset();

}

​

function loadKNNDataset() {

  if (isModelReady && isVideoReady) {

    //knnClassifier.load("./myKNNDataset.json", KNNDatasetReady);

  }

}

​

function KNNDatasetReady() {

  console.log("Loaded: KNN Dataset");

  updateCounts();

​

  initClassification();

}

​

function initClassification() {

  isClassifying = true;

  console.log("Init Classification");

  classify();

}

​

//INPUTS

​

function keyPressed() {

  //

  if (key === "0") {

    addExample("0");

  } else if (key === "1") {

    addExample("1");

  } else if (key === "2") {

    addExample("2");

  }

  // key press with shift

  else if (key === ")") {

    clearLabel("0");

  } else if (key === "!") {

    clearLabel("1");

  } else if (key === "@") {

    clearLabel("2");

  }

  // init the classification

  else if (key === " ") {

    classify();

  }

  // key press with shift

  else if (key === "S") {

    saveMyKNN();

  } else if (key === "L") {

    loadMyKNN();

  }

}

​

//--------------------------------------------------------------

//Modelling

//--------------------------------------------------------------

​

// Predict the current frame.

function classify() {

  // Get the total number of labels from knnClassifier

  const numLabels = knnClassifier.getNumLabels();

  if (numLabels <= 0) {

    console.error("There is no examples in any label");

    return;

  }

  // Get the features of the input video

  const features = featureExtractor.infer(capture);

  knnClassifier.classify(features, gotResults);

}

​

// Show the results

function gotResults(err, result) {

  // Display any error

  if (err) {

    console.error(err);

  }

​

  if (result.confidencesByLabel) {

    const confidences = result.confidencesByLabel;

    // result.label is the label that has the highest confidence

    if (result.label) {

      resultLabel = result.label;

      resultScore = confidences[result.label] * 100;

    }

    label0_score = confidences["mirrorSelfie"] * 100;

    label1_score = confidences["notMirrorSelfie"] * 100;

    //label2_score = confidences["2"] * 100;

  }

​

  classify();

}

​

function addExample(label) {

  const features = featureExtractor.infer(capture);

  knnClassifier.addExample(features, label);

  updateCounts();

}

​

// Clear the examples in one label

function clearLabel(label) {

  knnClassifier.clearLabel(label);

  updateCounts();

}

​

// Clear all the examples in all labels

function clearAllLabels() {

  knnClassifier.clearAllLabels();

  updateCounts();

}

​

// Save dataset as myKNNDataset.json

function saveMyKNN() {

  knnClassifier.save("myKNNDataset" + hour() + minute());

  console.log("Saved: KNN Dataset");

}

​

// Load dataset to the classifier

function loadMyKNN() {

  knnClassifier.load("./myKNNDataset.json", KNNDatasetReady);

}

​

function updateCounts() {

  const counts = knnClassifier.getCountByLabel();

  label0_count = counts["mirrorSelfie"];

  label1_count = counts["notMirrorSelfie"];

  // label2_count = counts["2"];

}

​

​