// https://editor.p5js.org/ml5/sketches/FeatureExtractor_Image_Regression

​

let featureExtractor;

let regressor;

let video;

let statusString = "STATUS_ADD_SAMPLES";

let trainingLoss;

let nSamples = 0;

let predictionValue01;

​

let windowOpen; //how open the window is

let roomFill; //how full the room is

let projectile;

let attacking;

let timeStart;

​

let flyingSound;

let splatSound;

let smashSound;

let music;

let static;

let hitSound;

let closefilter;

let paused;

​

let scene;

​

let ease = new p5.Ease();

​

function preload() {

  flyingSound = loadSound("Sounds/falling.wav");

  splatSound = loadSound("Sounds/splat.wav");

  smashSound = loadSound("Sounds/smash.mp3");

  music = loadSound("Sounds/music.wav");

  static = loadSound("Sounds/static.wav");

  hitSound = loadSound("Sounds/hit.wav");

}

​

function attack() {

  flyingSound.play();

  projectile = 0;

  attacking = true;

  timeStart = millis();

}

​

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

function setup() {

  createCanvas(320, 240);

  video = createCapture(VIDEO);

  video.hide();

​

  attacking = false;

  windowOpen = 0;

  closefilter = new p5.LowPass();

  closefilter.freq(20);

  flyingSound.disconnect();

  smashSound.disconnect();

  splatSound.disconnect();

  music.disconnect();

  flyingSound.connect(closefilter);

  smashSound.connect(closefilter);

  splatSound.connect(closefilter);

  music.connect(closefilter);

  paused = false;

​

  //music.loop();

  music.setVolume(0.05);

  splatSound.setVolume(3);

  hitSound.setVolume(3);

​

  static.loop();

​

  projectile = 0;

​

  roomFill = 0;

  scene = 1;

​

  featureExtractor = ml5.featureExtractor('MobileNet');

  regressor = featureExtractor.regression(video);

}

​

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

function draw() {

​

​

  windowOpen = map(mouseX, 0, width, 0, 1);

​

  static.setVolume(roomFill);

  // ml5 stuff -------------------------------

  background('red');

  tint(255, 255, 255, 64);

  image(video, 0, 0, 320, 240);

  /*

  // draw a very simple "slider"

  noFill();

  stroke(0);

  rect(0, 0, width, 10);

  var mx = mouseX;

  if (predictionValue01) {

    mx = map(predictionValue01, 0, 1, 0, width);

  }

  rect(mx, 0, 1, 10);

​

  // draw a red circle whose X location is

  // proportional to our predicted value

  var positionX = width / 2;

  if (predictionValue01) {

    positionX = map(predictionValue01, 0, 1, 0, width);

    windowOpen = predictionValue01;

  }

  noStroke();

  fill(255, 0, 0);

  ellipse(positionX, 120, 50, 50);

  */

  // draw diagnostic/debug information

  /*

  fill('black');

  var instructions = "Press s and adjust mouseX to add samples. \n";

  instructions += "Press t to train model. \n";

  instructions += "Press p to start predicting. \n";

  text(instructions, 15, 30);

​

  text("status: " + statusString, 15, 75);

  text("nSamples: " + nSamples, 15, 90);

  text("trainingLoss: " + trainingLoss, 15, 105);

  var pStr = (predictionValue01) ? nf(predictionValue01, 1, 3) : "undefined";

  text("prediction: " + pStr, 15, 120);

  */

  // my stuff ---------------------------------

​

  if (predictionValue01) {

​

    if (!attacking && random() < 0.008 && !paused) {

      attack();

    }

​

    if (attacking) {

      projectile = map(millis() - timeStart, 0, 3800, 0, 1);

    }

​

    if (attacking && projectile >= 1) {

      attacking = false;

      if (windowOpen < 0.3) {

        if (random() > 0.5)

          splatSound.play();

        else

          smashSound.play();

        flyingSound.stop();

      } else {

        hitSound.play();

        roomFill += 0.15;

      }

      projectile = 0;

    }

​

    let openness = ease.adjustableCenterDoubleExponentialSigmoid(windowOpen, 0.9, 0.2);

​

    closefilter.freq(map(openness, 0, 1, 830, 10000));

    flyingSound.setVolume(map(openness, 0, 1, 0.2, 1) * projectile);

    music.setVolume(map(openness, 0, 1, 0.0, 0.01));

    roomFill += map(openness, 0, 1, 0.00005, -0.0005);

​

    if (roomFill > 1)

      roomFill = 1;

    if (roomFill < 0)

      roomFill = 0;

​

    print(roomFill);

​

  }

  textAlign(CENTER);

  if (scene == 1)

  {

    tint(255, 255, 255, 64);

    text("point your webcam at a window",width/2,height/2);

    text("press Enter",width/2,height/2 + 20);

  }

  else if (scene == 2)

  {

    background(0,255,0);

    text("close the window all the way",width/2,height/2);

    text("then press Enter",width/2,height/2 + 20);

  }

  else if (scene == 3)

  {

    background(0,0,255);

    text("open the window all the way",width/2,height/2);

    text("then press Enter",width/2,height/2 + 20);

  }

  else if (scene == 4)

  {

    background(255,255,0);

    text("open the window halfway",width/2,height/2);

    text("then press Enter",width/2,height/2 + 20);

  }

  else if (scene == 6)

  {

    background(255,0,255);

    text("press Enter to begin",width/2,height/2);

  }

  else if (scene == 5)

  {

    background(0,255,255);

    text("loading...",width/2,height/2);

    if (nSamples == 6)

    {

      regressor.train(function(lossValue) {

        if (lossValue) {

          trainingLoss = lossValue;

          statusString = "STATUS_TRAINING_MODEL";

        } else {

          statusString = "STATUS_DONE_TRAINING";

        }

      });

      scene = 6;

    }

  }

  else if (scene == 7)

  {

    background(0,255,255);

    text("close your window to block oncoming baddies",width/2,height/2);

    text("open it to let the air out",width/2,height/2 + 20);

  }

}

​

​

​

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

function keyPressed() {

  if (keyCode == ENTER)

  {

    if (scene == 1) {

      scene = 2;

    } else if (scene == 2) {

      regressor.addImage(0);

      nSamples++;

      regressor.addImage(0);

      nSamples++;

      scene = 3;

    } else if (scene == 3) {

      regressor.addImage(1);

      nSamples++;

      regressor.addImage(1);

      nSamples++;

      scene = 4;

    } else if (scene == 4) {

      regressor.addImage(0.5);

      nSamples++;

      regressor.addImage(0.5);

      nSamples++;

      scene = 5;      

    } else if (scene == 6) {

      statusString = "STATUS_PREDICTING";

      regressor.predict(gotResultsCallback);

      scene = 7;

    }

  }

  if (key == 's') { // add sample

    var xValue = constrain(map(mouseX, 0, width, 0, 1), 0, 1);

    regressor.addImage(xValue);

    nSamples++;

​

  } else if (key == 't') { // train the regressor

    regressor.train(function(lossValue) {

      if (lossValue) {

        trainingLoss = lossValue;

        statusString = "STATUS_TRAINING_MODEL";

      } else {

        statusString = "STATUS_DONE_TRAINING";

      }

    });

​

  } else if (key == 'p') { // initiate prediction

    statusString = "STATUS_PREDICTING";

    regressor.predict(gotResultsCallback);

  } else if (key == 'b') {

    paused = !paused;

  }

}

​

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

// Store the results, and restart the process.

function gotResultsCallback(err, result) {

  if (err) {

    console.error(err);

  }

  if (result && result.value) {

    predictionValue01 = result.value;

    regressor.predict(gotResultsCallback);

  }

}