// Cloned by robby on 19 Nov 2022 from World "Character recognition neural network" by "Coding Train" project
// Please leave this clone trail here.
// Port of Character recognition neural network from here:
// https://github.com/CodingTrain/Toy-Neural-Network-JS/tree/master/examples/mnist
// with many modifications
// --- defined by MNIST - do not change these ---------------------------------------
const PIXELS = 28; // images in data set are tiny
const PIXELS_X = 28; // Rohan: change image height
const PIXELS_Y = 28; // Rohan: change image width
const PIXELSSQUARED = PIXELS * PIXELS;
const DIM = 1; // Rohan: change 1 for grayscale and 3 for RGB
// number of training and test exemplars in the data set:
const NOTRAIN = 124800; // change if needed
const NOTEST = 20800; // change if needed
//--- can modify all these --------------------------------------------------
// no of nodes in network
const noinput = PIXELSSQUARED;
const nohidden = 64;
const nooutput = 10;
const learningrate = 0.1; // default 0.1
// should we train every timestep or not
let do_training = true;
// how many to train and test per timestep
const TRAINPERSTEP = 15;
const TESTPERSTEP = 5;
// multiply it by this to magnify for display
const ZOOMFACTOR = 7;
const ZOOMPIXELS = ZOOMFACTOR * PIXELS;
// 3 rows of
// large image + 50 gap + small image
// 50 gap between rows
const canvaswidth = ( PIXELS + ZOOMPIXELS ) + 50;
const canvasheight = ( ZOOMPIXELS * 3 ) + 100;
const DOODLE_THICK = 18; // thickness of doodle lines
const DOODLE_BLUR = 3; // blur factor applied to doodles
let mnist;
// all data is loaded into this
// mnist.train_images
// mnist.train_labels
// mnist.test_images
// mnist.test_labels
//Rohan: Define variable
let cnn;
let TrainingCnn;
let cnnMod;
//Rohan: Character or alhpabet which need to be recognised
let alpha =["a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m", "n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z"]
//let alpha = ["A", "B", "C", "D", "E", "F", "G", "H", "I", "J", "K", "L", "M", "N", "O", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", "Z"];
let trainrun = 1;
let train_index = 0;
let testrun = 1;
let test_index = 0;
let total_tests = 0;
let total_correct = 0;
// images in LHS:
let doodle, demo;
let doodle_exists = false;
let demo_exists = false;
let mousedrag = false; // are we in the middle of a mouse drag drawing?
// save inputs to global var to inspect
// type these names in console
var train_inputs, test_inputs, demo_inputs, doodle_inputs;
// Matrix.randomize() is changed to point to this. Must be defined by user of Matrix.
function randomWeight()
{
return ( AB.randomFloatAtoB ( -0.5, 0.5 ) );
// Coding Train default is -1 to 1
}
// make run header bigger
AB.headerCSS ( { "max-height": "95vh" } );
//--- start of AB.msgs structure: ---------------------------------------------------------
// We output a serious of AB.msgs to put data at various places in the run header
var thehtml;
// 1 Doodle header
//Rohan: clear and Redrawing doodle
thehtml = "<hr> <h1> 1. Doodle </h1> Top row: Doodle (left) and shrunk (right). <br> " +
" Draw your doodle in top LHS. <button onclick='cleanDoodle();' class='normbutton' >Clear and Redo</button> <br> " ;
AB.msg ( thehtml, 1 );
// 2 Doodle variable data (guess)
// 3 Training header
//Rohan: Pausing and Restarting the process of training
thehtml = "<hr> <h1> 2. Training </h1> Middle row: Training image magnified (left) and original (right). <br> " +
" <button onclick='do_training = false;' class='normbutton' >Pause training</button> " +
" <button onclick='do_training = true;' class='normbutton' >Resume training</button> <br> ";
AB.msg ( thehtml, 3 );
// 4 variable training data
// 5 Testing header
thehtml = "<h3> Hidden tests </h3> " ;
AB.msg ( thehtml, 5 );
// 6 variable testing data
// 7 Demo header
thehtml = "<hr> <h1> 3. Demo </h1> Bottom row: Test image magnified (left) and original (right). <br>" +
" The network is <i>not</i> trained on any of these images. <br> " +
" <button onclick='makeDemo();' class='normbutton' >Demo test image</button> <br> ";
AB.msg ( thehtml, 7 );
// 8 Demo variable data (random demo ID)
// 9 Demo variable data (changing guess)
const bluespan = "<span style='font-weight:bold; font-size:large; color:darkblue'> " ;
//--- end of AB.msgs structure: ---------------------------------------------------------
function setup()
{
createCanvas ( canvaswidth, canvasheight );
doodle = createGraphics ( ZOOMPIXELS, ZOOMPIXELS ); // doodle on larger canvas
doodle.pixelDensity(1);
//Rohan: clean the doodle by using the below function to draw
cleanDoodle();
// JS load other JS
// maybe have a loading screen while loading the JS and the data set
AB.loadingScreen();
$.getScript ( "/uploads/codingtrain/matrix.js", function()
{
//Rohan: upload the convolutional neural netowrk which is to be used in building the model
//Ref:https://cs.stanford.edu/people/karpathy/convnetjs
$.getScript ( "/uploads/rohanb456/convnet.js", function()
{
//Rohan: upload the dataset
$.getScript ( "/uploads/rohanb456/mnist.js", function()
{
console.log ("All JS Files loaded");
//Rohan: Implement Simple CNN with 1 input layer, 3 convo and pooling layer and 1 output layer
let layers = [];
//-------input layer------
//Rohan: input(ip) layer with 28*28 pixels and is in grayscale(out_depth:1) if color then out_depth:3
layers.push({type:"input", out_sx:PIXELS_X, out_sy:PIXELS_Y, out_depth:DIM});
//Rohan: output(op) Volume is of size 28x28x1
//------hidden layer------
//first layer
//can use different activation functions like relu,sigmoid,softmax etc
layers.push({type:"conv", sx:5, filters:8, stride:1, pad:2, activation:"tanh"});
//Rohan: convolutional layer that has 8 kernels, and size 5*5 along with 2 pixels padded on all the sides to make output of same size. Also have tanh Activation fuction.
//Rohan: output(op) Volume is of size 28x28x8
layers.push({type:"pool", sx:2, stride:2});
//Rohan: Pooling layer that has 2 strides, and size of 2*2
//Rohan: output(op) Volume is of size 14x14x8
//second layer
layers.push({type:"conv", sx:5, filters:16, stride:1, pad:2, activation:"tanh"});
//Rohan: convolutional layer that has 16 kernels, and size 5*5 along with 2 pixels padded on all the sides to make output of same size. Also have tanh Activation fuction.
//Rohan: output(op) Volume is of size 14x14x16
layers.push({type:"pool", sx:2, stride:2});
//Rohan: Pooling layer that has 2 strides, and size of 2*2
//third layer
layers.push({type:"conv", sx:5, filters:20, stride:1, pad:2, activation:"tanh"});
//Rohan: convolutional layer that has 20 kernels, and size 5*5 along with 2 pixels padded on all the sides to make output of same size. Also have tanh Activation fuction.
//Rohan: output(op) Volume is of size 7x7x20
layers.push({type:"pool", sx:2, stride:2});
//Rohan: Pooling layer that has 2 strides, and size of 2*2
//Rohan: output(op) Volume is of size 3x3x20
//layers.push({type:'fc', num_neurons:20, activation:'tanh'});
//------output layer-------
layers.push({type:"softmax", num_classes:26});
//Rohan: output(op) layer whose Volume is of size 1x1x26 along with softmax activation function as we dealing with more than two class or character recogntion
/*
//Rohan: Tried to implement LeNet-5 but failed mainly due to high processing time requirement as the model is built to work on 32*32 size data
layers.push({type : "input",
out_sx : 32,
out_sy : 32,
out_depth : 1});
layers.push({type : "conv",
sx : 5,
filters : 6,
stride : 1,
activation : "tanh"});
layers.push({type : "pool",
sx : 2,
stride : 2});
layers.push({type : "conv",
sx : 5,
filters : 16,
stride : 1,
activation : "tanh"});
layers.push({type : "pool",
sx : 2,
stride : 2});
layers.push({type : "conv",
sx : 5,
filters : 120,
stride : 1,
activation : "tanh"});
layers.push({type:'fc', activation:'tanh'});
layers.push({type:'fc', activation:'tanh'});
layers.push({type:'softmax', num_classes:26});
*/
//Rohan: build a simple netowrk using convnetjs
cnnMod = new convnetjs.Net();
//var scores = net.forward(cnnMod); // pass forward through network
// to check the score
//console.log('score for class 0 is assigned:' + scores.w[0]);
//Rohan: specifing the layers in the cnnMod
cnnMod.makeLayers(layers);
//Rohan: Stocastic Gradient Descent Trainer is used with momentum 0.9, 5 batch size and l2_decay 0.0001
//rohan: add learning_rate parameter by replacing method parameter i.e learning_rate:0.1,0.001,0.0001
TrainingCnn = new convnetjs.SGDTrainer(cnnMod, {method : "adadelta", momentum: 0.9, batch_size : 5, l2_decay : 0.0001});
//Rohan: Calling the loadData function
loadData();
});
});
});
}
// load data set from local file (on this server)
function loadData()
{
loadMNIST (function(data){
mnist = data;
//Rohan: rotate the images
let index = 0;
for (; index < NOTRAIN; index++) {
imageRotation(mnist.train_images[index]);
}
for (index = 0; index < NOTEST; index++) {
imageRotation(mnist.test_images[index]);
}
console.log ("All data loaded into Emnist object.")
console.log(mnist);
AB.removeLoading(); // if no loading screen exists, this does nothing
});
}
function getImage ( img ) // make a P5 image object from a raw data array
{
let theimage = createImage (PIXELS, PIXELS); // make blank image, then populate it
theimage.loadPixels();
for (let i = 0; i < PIXELSSQUARED ; i++)
{
let bright = img[i];
let index = i * 4;
theimage.pixels[index + 0] = bright;
theimage.pixels[index + 1] = bright;
theimage.pixels[index + 2] = bright;
theimage.pixels[index + 3] = 255;
}
theimage.updatePixels();
return theimage;
}
function getInputs ( img ) // convert img array into normalised input array
{
let inputs = [];
for (let i = 0; i < PIXELSSQUARED ; i++)
{
let bright = img[i];
inputs[i] = bright / 255; // normalise to 0 to 1
}
return ( inputs );
}
//Rohan: below function rotates the images which helps to build a more reliable model
function imageRotation(img) {
for (let e = 0; e < PIXELS; e++) {
for (let f = e; f < PIXELS; f++) {
let a = e * PIXELS + f;
let b = f * PIXELS + e;
let c = img[a];
img[a] = img[b];
img[b] = c;
}
}
}
function trainit (show) // train the network with a single exemplar, from global var "train_index", show visual on or off
{
let img = mnist.train_images[train_index];
let label = mnist.train_labels[train_index];
// optional - show visual of the image
if (show)
{
var theimage = getImage ( img ); // get image from data array
image ( theimage, 0, ZOOMPIXELS+50, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimage, ZOOMPIXELS+50, ZOOMPIXELS+50, PIXELS, PIXELS ); // original
}
// set up the inputs
let inputs = getInputs ( img ); // get inputs from data array
train_inputs = inputs;
{
//Rohan: fetchcnnip creates input activations required for trainer.
let set = fetchcnnip(inputs);
//Rohan: pass exemplar and their labels for classification
TrainingCnn.train(set, label);
}
// can inspect in console
thehtml = " trainrun: " + trainrun + "<br> no: " + train_index ;
AB.msg ( thehtml, 4 );
train_index++;
if ( train_index == NOTRAIN )
{
train_index = 0;
console.log( "finished trainrun: " + trainrun );
trainrun++;
}
}
//Rohan: creating a volume of inp activations with 28*28 size and with 0 intialization.
function fetchcnnip(t) {
var x = new convnetjs.Vol(28, 28, 1, 0);
for (var j = 0; j < PIXELSSQUARED; j++) {
x.w[j] = t[j];
}
return x;
}
function testit() // test the network with a single exemplar, from global var "test_index"
{
let img = mnist.test_images[test_index];
let label = mnist.test_labels[test_index];
// set up the inputs
let inputs = getInputs ( img );
let cnninput = fetchcnnip(inputs);
//get the test image and find maximum weight
test_inputs = inputs; // can inspect in console
let place = findMax(cnnMod.forward(cnninput).w);
var img1 = getImage(img);
image(img1, 0, ZOOMPIXELS + 50, ZOOMPIXELS, ZOOMPIXELS);
image(img1, ZOOMPIXELS + 50, ZOOMPIXELS + 50, PIXELS, PIXELS);
total_tests++;
if (place == label) total_correct++;
let percent = (total_correct / total_tests) * 100 ;
thehtml = " testrun: " + testrun + "<br> no: " + total_tests + " <br> " +
" correct: " + total_correct + "<br>" +
" score: " + bluespan + percent.toFixed(2) + "</span>";
AB.msg ( thehtml, 6 );
test_index++;
if ( test_index == NOTEST )
{
console.log( "finished testrun: " + testrun + " score: " + percent.toFixed(2) );
testrun++;
test_index = 0;
total_tests = 0;
total_correct = 0;
}
}
//--- find no.1 (and maybe no.2) output nodes ---------------------------------------
// (restriction) assumes array values start at 0 (which is true for output nodes)
function find12 (a) // return array showing indexes of no.1 and no.2 values in array
{
let no1 = 0;
let no2 = 0;
let no1value = 0;
let no2value = 0;
for (let i = 0; i < a.length; i++)
{
if (a[i] > no1value) // new no1
{
// old no1 becomes no2
no2 = no1;
no2value = no1value;
// now put in the new no1
no1 = i;
no1value = a[i];
}
else if (a[i] > no2value) // new no2
{
no2 = i;
no2value = a[i];
}
}
var b = [ no1, no2 ];
return b;
}
// just get the maximum - separate function for speed - done many times
// find our guess - the max of the output nodes array
function findMax (a)
{
let no1 = 0;
let no1value = 0;
for (let i = 0; i < a.length; i++)
{
if (a[i] > no1value)
{
no1 = i;
no1value = a[i];
}
}
return no1;
}
// --- the draw function -------------------------------------------------------------
// every step:
function draw()
{
// check if libraries and data loaded yet:
if ( typeof mnist == 'undefined' ) return;
background ('black');
strokeWeight(1);
stroke('blue');
rect(0,0,ZOOMPIXELS,ZOOMPIXELS);
textSize(15);
textAlign(CENTER);
text("Please Draw",ZOOMPIXELS/2,ZOOMPIXELS/2);
if ( do_training )
{
// do some training per step
for (let i = 0; i < TRAINPERSTEP; i++)
{
if (i === 0) trainit(true); // show only one per step - still flashes by
else trainit(false);
}
// do some testing per step
for (let i = 0; i < TESTPERSTEP; i++)
testit();
}
// keep drawing demo and doodle images
// and keep guessing - we will update our guess as time goes on
if ( demo_exists )
{
drawDemo();
guessDemo();
}
if ( doodle_exists )
{
drawDoodle();
guessDoodle();
}
// detect doodle drawing
// (restriction) the following assumes doodle starts at 0,0
if ( mouseIsPressed ) // gets called when we click buttons, as well as if in doodle corner
{
// console.log ( mouseX + " " + mouseY + " " + pmouseX + " " + pmouseY );
var MAX = ZOOMPIXELS + 20; // can draw up to this pixels in corner
if ( (mouseX < MAX) && (mouseY < MAX) && (pmouseX < MAX) && (pmouseY < MAX) )
{
mousedrag = true; // start a mouse drag
doodle_exists = true;
doodle.stroke('red');
strokeJoin(ROUND);
doodle.strokeWeight( DOODLE_THICK );
doodle.line(mouseX, mouseY, pmouseX, pmouseY);
}
}
else
{
// are we exiting a drawing
if ( mousedrag )
{
mousedrag = false;
// console.log ("Exiting draw. Now blurring.");
doodle.filter (BLUR, DOODLE_BLUR); // just blur once
// console.log (doodle);
}
}
}
//--- demo -------------------------------------------------------------
// demo some test image and predict it
// get it from test set so have not used it in training
function makeDemo()
{
demo_exists = true;
var i = AB.randomIntAtoB ( 0, NOTEST - 1 );
demo = mnist.test_images[i];
var label = mnist.test_labels[i];
thehtml = "Test image no: " + i + "<br>" +
"Classification: " + alpha[label - 1] + "<br>" ;
AB.msg ( thehtml, 8 );
// type "demo" in console to see raw data
}
function drawDemo()
{
var theimage = getImage ( demo );
// console.log (theimage);
image ( theimage, 0, canvasheight - ZOOMPIXELS, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimage, ZOOMPIXELS+50, canvasheight - ZOOMPIXELS, PIXELS, PIXELS ); // original
}
function guessDemo()
{
let inputs = getInputs ( demo );
demo_inputs = inputs; // can inspect in console
//Rohan: fetch input and find the max weight
let cnninput = fetchcnnip(inputs);
let j = findMax(cnnMod.forward(cnninput).w);
thehtml = " We classify it as: " + bluespan + alpha[j-1] + "</span>" ;
AB.msg ( thehtml, 9 );
}
//--- doodle -------------------------------------------------------------
function drawDoodle()
{
// doodle is createGraphics not createImage
let theimage = doodle.get();
// console.log (theimage);
image ( theimage, 0, 0, ZOOMPIXELS, ZOOMPIXELS ); // original
image ( theimage, ZOOMPIXELS+20, 0, PIXELS, PIXELS ); // shrunk
}
function guessDoodle()
{
// doodle is createGraphics not createImage
let img = doodle.get();
img.resize ( PIXELS, PIXELS );
img.loadPixels();
// set up inputs
let inputs = [];
for (let i = 0; i < PIXELSSQUARED ; i++)
{
inputs[i] = img.pixels[i * 4] / 255;
}
doodle_inputs = inputs; // can inspect in console
//Rohan: shows two most possible predictions
let cnninput = fetchcnnip(inputs);
let a = find12(cnnMod.forward(cnninput).w);
thehtml = " Our 1st Guess is: " + bluespan + alpha[a[0] - 1] + "</span> <br>" +
" Our 2nd Guess is: " + bluespan + alpha[a[1] - 1] + "</span>";
AB.msg ( thehtml, 2 );
}
function cleanDoodle()
{
doodle_exists = false;
doodle.background('black');
}
// --- debugging --------------------------------------------------
// in console
// showInputs(demo_inputs);
// showInputs(doodle_inputs);
function showInputs ( inputs )
// display inputs row by row, corresponding to square of pixels
{
var str = "";
for (let i = 0; i < inputs.length; i++)
{
if ( i % PIXELS === 0 ) str = str + "\n"; // new line for each row of pixels
var value = inputs[i];
str = str + " " + value.toFixed(2) ;
}
console.log (str);
}