// Cloned by Prateek Bhargava on 11 Dec 2020 from World "Character recognition neural network" by "Coding Train" project
// Please leave this clone trail here.
// ====================================================================================================================
// This code is designed for use on the Ancient Brain site.
// This code is built as a part of Assignment 2 of CA686 course
// The Base code is taken from World "Character recognition neural network" hosted on Ancient Brian;
// and modifications were performed on top of it
// ====================================================================================================================
// ==== Modifications =================================================================================================
// Modifications performed by: Prateek Bhargava [student ID: 20210628]
// Modification details: Data augmentation to enhance the performance of Doodle recognition
// Training dataset of 60000 images enhanced 5 times
// by adding instances of the same images with shifted pixels: shfting the image right, left, up, down
// ====================================================================================================================
// Port of Character recognition neural network from here:
// https://github.com/CodingTrain/Toy-Neural-Network-JS/tree/master/examples/mnist
// --- defined by MNIST - do not change these ---------------------------------------
const PIXELS = 28; // images in data set are tiny
const PIXELSSQUARED = PIXELS * PIXELS;
/* Code by Prateek ---- Begin
Explanation: SHIFT_FACTOR defines how much we wish to shift the images (by how many pixels sideways)
VERTICAL_SHIFT_FACTOR calculates the number of pixels to be shifted, in order to achieve vertical shift
of the same factor as SHIFT_FACTOR.
*/
const SHIFT_FACTOR = 5 // number of pixel lengths to be shifted
const VERTICAL_SHIFT_FACTOR = SHIFT_FACTOR * PIXELS // SHIFT_FACTOR * 28
const PIXELSSQUARED_H_SHIFT = PIXELSSQUARED - SHIFT_FACTOR
const PIXELSSQUARED_V_SHIFT = PIXELSSQUARED - VERTICAL_SHIFT_FACTOR
/* Code by Prateek ---- End */
// number of training and test exemplars in the data set:
const NOTRAIN = 60000;
const NOTEST = 10000;
//--- 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;
/* Code by Prateek ---- Begin
Explanation: Since we are performing model training 5 times in each step,
following values need to be changed accordingly.
*/
// how many to train and test per timestep
const TRAINPERSTEP = 18;
const TESTPERSTEP = 3;
/* Code by Prateek ---- End */
// 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) * 5;
const canvasheight = ( ZOOMPIXELS * 3 ) + 100;
const DOODLE_THICK = 18; // thickness of doodle lines
const DOODLE_BLUR = 3; // blur factor applied to doodles
/* Code by Prateek ---- Begin */
const DOODLE_POSTERIZE = 4;
/* Code by Prateek ---- End */
let mnist;
// all data is loaded into this
// mnist.train_images
// mnist.train_labels
// mnist.test_images
// mnist.test_labels
let nn;
let trainrun = 1;
let train_index = 0;
let total_train_index = 0; //total including the shifted images (left/right/up/down)
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
}
// CSS trick
// make run header bigger
$("#runheaderbox").css ( { "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
thehtml = "<hr> <h1> 1. Doodle </h1> Top row: Doodle (left) and shrunk (right). <br> " +
" Draw your doodle in top LHS. <button onclick='wipeDoodle();' class='normbutton' >Clear doodle</button> <br> ";
AB.msg ( thehtml, 1 );
// 2 Doodle variable data (guess)
// 3 Training header
thehtml = "<hr> <h1> 2. Training </h1> Middle row: Training image magnified (left) and original (right). <br> " +
" <button onclick='do_training = false;' class='normbutton' >Stop 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 greenspan = "<span style='font-weight:bold; font-size:x-large; color:darkgreen'> " ;
//--- end of AB.msgs structure: ---------------------------------------------------------
function setup()
{
createCanvas ( canvaswidth, canvasheight );
textSize(20);
doodle = createGraphics ( ZOOMPIXELS, ZOOMPIXELS ); // doodle on larger canvas
doodle.pixelDensity(1);
// loading screen while loading the JS and the data set
AB.loadingScreen();
$.getScript ( "/uploads/codingtrain/matrix.js", function()
{
$.getScript ( "/uploads/codingtrain/nn.js", function()
{
$.getScript ( "/uploads/codingtrain/mnist.js", function()
{
console.log ("All JS loaded");
nn = new NeuralNetwork( noinput, nohidden, nooutput );
nn.setLearningRate ( learningrate );
loadData();
});
});
});
}
// load data set from local file (on this server)
function loadData()
{
loadMNIST ( function(data)
{
mnist = data;
console.log ("All data loaded into mnist 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 );
}
/* Code by Prateek ---- Begin
Explanation: Data Augmentation. Creating new training data by augmenting available data.
*/
/* Right Shift
inputs_R_SHIFT holds 'normalized pixel brightness' information of the same image as original but
shifted towards right, by the margin of SHIFT_FACTOR pixels.
Because of this right-shift, first few pixels will be empty (in our case, black).
And we loose information of last few pixels from the original image.
*/
function getInputs_R_SHIFT ( img ) // convert img array into normalised input array
{
let inputs_R_SHIFT = [];
for (let k = 0; k < SHIFT_FACTOR ; k++) inputs_R_SHIFT[k] = 0; // Fill the first few pixels with 0
for (let i = 0; i < PIXELSSQUARED_H_SHIFT ; i++) //PIXELSSQUARED_H_SHIFT = PIXELSSQUARED - SHIFT_FACTOR
{
let bright = img[i];
inputs_R_SHIFT[i+SHIFT_FACTOR] = bright / 255; // normalise to 0 to 1
}
return ( inputs_R_SHIFT );
}
/* Left Shift
inputs_L_SHIFT holds 'normalized pixel brightness' information of the same image as original but
shifted towards left, by the margin of SHIFT_FACTOR pixels.
Because of this left-shift, last few pixels will be empty (in our case, black).
And we loose information of first few pixels from the original image.
*/
function getInputs_L_SHIFT ( img )
{
let inputs_L_SHIFT = [];
for (let i = SHIFT_FACTOR; i < PIXELSSQUARED ; i++)
{
let bright = img[i];
inputs_L_SHIFT[i-SHIFT_FACTOR] = bright / 255;
}
for (let k = PIXELSSQUARED_H_SHIFT; k < PIXELSSQUARED ; k++) inputs_L_SHIFT[k] = 0; // Fill the last few pixels with 0
return ( inputs_L_SHIFT );
}
/* Up Shift
inputs_UP_SHIFT holds 'normalized pixel brightness' information of the same image as original but
shifted upside, by the margin of SHIFT_FACTOR pixels.
Because of this up-shift, last few pixels will be empty (in our case, black).
And we loose information of first few pixels from the original image.
*/
function getInputs_UP_SHIFT ( img )
{
let inputs_UP_SHIFT = [];
for (let i = VERTICAL_SHIFT_FACTOR; i < PIXELSSQUARED ; i++)
{
let bright = img[i];
inputs_UP_SHIFT[i-VERTICAL_SHIFT_FACTOR] = bright / 255;
}
for (let k = PIXELSSQUARED_V_SHIFT; k < PIXELSSQUARED ; k++) inputs_UP_SHIFT[k] = 0;
return ( inputs_UP_SHIFT );
}
/* Down Shift
inputs_DN_SHIFT holds 'normalized pixel brightness' information of the same image as original but
shifted downside, by the margin of SHIFT_FACTOR pixels.
Because of this down-shift, first few pixels will be empty (in our case, black).
And we loose information of last few pixels from the original image.
*/
function getInputs_DN_SHIFT ( img )
{
let inputs_DN_SHIFT = [];
for (let k = 0; k < VERTICAL_SHIFT_FACTOR ; k++) inputs_DN_SHIFT[k] = 0;
for (let i = 0; i < PIXELSSQUARED_V_SHIFT ; i++) // PIXELSSQUARED_V_SHIFT= PIXELSSQUARED - VERTICAL_SHIFT_FACTOR
{
let bright = img[i];
inputs_DN_SHIFT[i+VERTICAL_SHIFT_FACTOR] = bright / 255;
}
return ( inputs_DN_SHIFT );
}
/* Code by Prateek ---- Begin
Explanation: Un-normalized values, used to display imgaes on the canvas.
*/
function getInputs_R_SHIFT_show ( img ) // convert img array into normalised input array
{
let inputs_R_SHIFT_show = [];
for (let k = 0; k < SHIFT_FACTOR ; k++) inputs_R_SHIFT_show[k] = 0; // Fill the first few pixels with 0
for (let i = 0; i < PIXELSSQUARED_H_SHIFT ; i++) //PIXELSSQUARED_H_SHIFT = PIXELSSQUARED - SHIFT_FACTOR
{
let bright = img[i];
inputs_R_SHIFT_show[i+SHIFT_FACTOR] = bright; // normalise to 0 to 1
}
return ( inputs_R_SHIFT_show );
}
function getInputs_L_SHIFT_show ( img )
{
let inputs_L_SHIFT_show = [];
for (let i = SHIFT_FACTOR; i < PIXELSSQUARED ; i++)
{
let bright = img[i];
inputs_L_SHIFT_show[i-SHIFT_FACTOR] = bright ;
}
for (let k = PIXELSSQUARED_H_SHIFT; k < PIXELSSQUARED ; k++) inputs_L_SHIFT_show[k] = 0; // Fill the last few pixels with 0
return ( inputs_L_SHIFT_show );
}
function getInputs_UP_SHIFT_show ( img )
{
let inputs_UP_SHIFT_show = [];
for (let i = VERTICAL_SHIFT_FACTOR; i < PIXELSSQUARED ; i++)
{
let bright = img[i];
inputs_UP_SHIFT_show[i-VERTICAL_SHIFT_FACTOR] = bright ;
}
for (let k = PIXELSSQUARED_V_SHIFT; k < PIXELSSQUARED ; k++) inputs_UP_SHIFT_show[k] = 0;
return ( inputs_UP_SHIFT_show );
}
function getInputs_DN_SHIFT_show ( img )
{
let inputs_DN_SHIFT_show = [];
for (let k = 0; k < VERTICAL_SHIFT_FACTOR ; k++) inputs_DN_SHIFT_show[k] = 0;
for (let i = 0; i < PIXELSSQUARED_V_SHIFT ; i++) // PIXELSSQUARED_V_SHIFT= PIXELSSQUARED - VERTICAL_SHIFT_FACTOR
{
let bright = img[i];
inputs_DN_SHIFT_show[i+VERTICAL_SHIFT_FACTOR] = bright ;
}
return ( inputs_DN_SHIFT_show );
}
/* Code by Prateek ---- End */
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];
/* Code by Prateek ---- Begin
*/
let inputs_R_SHIFT_show = getInputs_R_SHIFT_show ( img );
let inputs_L_SHIFT_show = getInputs_L_SHIFT_show ( img );
let inputs_UP_SHIFT_show = getInputs_UP_SHIFT_show ( img );
let inputs_DN_SHIFT_show = getInputs_DN_SHIFT_show ( img );
/* Code by Prateek ---- End */
/* Code by Prateek ---- Begin
*/
if (show)
{
var theimage = getImage ( img ); // get image from data array
image ( theimage, 25, ZOOMPIXELS+50, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimage, ZOOMPIXELS+45, ZOOMPIXELS+50, PIXELS, PIXELS ); // original
var theimageR = getImage ( inputs_R_SHIFT_show );
image ( theimageR, ZOOMPIXELS+90, ZOOMPIXELS+50, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimageR, ZOOMPIXELS+ZOOMPIXELS+110, ZOOMPIXELS+50, PIXELS, PIXELS ); // original
var theimageL = getImage ( inputs_L_SHIFT_show );
image ( theimageL, ZOOMPIXELS+ZOOMPIXELS+155, ZOOMPIXELS+50, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimageL, ZOOMPIXELS+ZOOMPIXELS+ZOOMPIXELS+110+65, ZOOMPIXELS+50, PIXELS, PIXELS ); // original
var theimageUP = getImage ( inputs_UP_SHIFT_show );
image ( theimageUP, ZOOMPIXELS+ZOOMPIXELS+155+ZOOMPIXELS+65, ZOOMPIXELS+50, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimageUP, ZOOMPIXELS+ZOOMPIXELS+ZOOMPIXELS+110+65+ZOOMPIXELS+65, ZOOMPIXELS+50, PIXELS, PIXELS ); // original
var theimageDN = getImage ( inputs_DN_SHIFT_show );
image ( theimageDN, ZOOMPIXELS+ZOOMPIXELS+155+ZOOMPIXELS+65+ZOOMPIXELS+65, ZOOMPIXELS+50, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimageDN, ZOOMPIXELS+ZOOMPIXELS+ZOOMPIXELS+110+65+ZOOMPIXELS+65+ZOOMPIXELS+65, ZOOMPIXELS+50, PIXELS, PIXELS ); // original
}
// set up the inputs
let inputs = getInputs ( img ); // get inputs from data array
/* Code by Prateek ---- Begin
*/
let inputs_R_SHIFT = getInputs_R_SHIFT ( img );
let inputs_L_SHIFT = getInputs_L_SHIFT ( img );
let inputs_UP_SHIFT = getInputs_UP_SHIFT ( img );
let inputs_DN_SHIFT = getInputs_DN_SHIFT ( img );
/* Code by Prateek ---- End */
// set up the outputs
let targets = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0];
targets[label] = 1; // change one output location to 1, the rest stay at 0
// console.log(train_index);
// console.log(inputs);
// console.log(targets);
train_inputs = inputs; // can inspect in console
nn.train ( inputs, targets );
//nn.train ( inputs, targets );
/* Code by Prateek ---- Begin
Explanation: Training the model with the new set of data
*/
nn.train ( inputs_R_SHIFT, targets );
nn.train ( inputs_L_SHIFT, targets );
nn.train ( inputs_UP_SHIFT, targets );
nn.train ( inputs_DN_SHIFT, targets );
/* Code by Prateek ---- End */
total_train_index = train_index * 5;
thehtml = " Trainrun: " + trainrun + "<br> Total: " + total_train_index + "<br> Original: " + train_index
+ "<br> Right_shifted: " + train_index + "<br> Left_shifted: " + train_index + "<br> Up_shifted: " + train_index
+ "<br> Down_shifted: " + train_index + "<br> [Shifted by: " + SHIFT_FACTOR + " pixels]" ;
AB.msg ( thehtml, 4 );
train_index++;
if ( train_index == NOTRAIN )
{
train_index = 0;
console.log( "finished trainrun: " + trainrun );
trainrun++;
}
}
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 );
test_inputs = inputs; // can inspect in console
let prediction = nn.predict(inputs); // array of outputs
let guess = findMax(prediction); // the top output
total_tests++;
if (guess == label) total_correct++;
let percent = (total_correct / total_tests) * 100 ;
thehtml = " testrun: " + testrun + "<br> no: " + total_tests + " <br> " +
" correct: " + total_correct + "<br>" +
" score: " + greenspan + 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)
{
no1 = i;
no1value = a[i];
}
else if (a[i] > no2value)
{
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;
/* Code by Prateek ---- Begin */
background ('#F4D03F');
rect(25, ZOOMPIXELS + 1, ZOOMPIXELS + 1, 2);
rect(25, 2 * ZOOMPIXELS + 80, ZOOMPIXELS + 2, 2);
text('Original', 75, ZOOMPIXELS + 25);
text('Right Shifted', 50+ZOOMPIXELS+90, ZOOMPIXELS + 25);
text('Left Shifted', 50+ZOOMPIXELS+ZOOMPIXELS+155, ZOOMPIXELS + 25);
text('Up Shifted', 50+ZOOMPIXELS+ZOOMPIXELS+155+ZOOMPIXELS+65, ZOOMPIXELS + 25);
text('Down Shifted', 45+ZOOMPIXELS+ZOOMPIXELS+155+ZOOMPIXELS+65+ZOOMPIXELS+65, ZOOMPIXELS + 25);
/* Code by Prateek ---- End */
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 //PRATEEK
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('white');
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
doodle.filter(POSTERIZE, DOODLE_POSTERIZE);
// 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: " + label + "<br>" ;
AB.msg ( thehtml, 8 );
// type "demo" in console to see raw data
}
function drawDemo()
{
var theimage = getImage ( demo );
// console.log (theimage);
image ( theimage, 25, canvasheight - ZOOMPIXELS, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimage, ZOOMPIXELS+45, canvasheight - ZOOMPIXELS, PIXELS, PIXELS ); // original
}
function guessDemo()
{
let inputs = getInputs ( demo );
demo_inputs = inputs; // can inspect in console
let prediction = nn.predict(inputs); // array of outputs
let guess = findMax(prediction); // the top output
thehtml = " We classify it as: " + greenspan + guess + "</span>" ;
AB.msg ( thehtml, 9 );
}
//--- doodle -------------------------------------------------------------
function drawDoodle()
{
// doodle is createGraphics not createImage
let theimage = doodle.get();
// console.log (theimage);
image ( theimage, 25, 0, ZOOMPIXELS, ZOOMPIXELS ); // original
image ( theimage, ZOOMPIXELS+45, 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
// feed forward to make prediction
let prediction = nn.predict(inputs); // array of outputs
let b = find12(prediction); // get no.1 and no.2 guesses
thehtml = " We classify it as: " + greenspan + b[0] + "</span> <br>" +
" No.2 guess is: " + greenspan + b[1] + "</span>";
AB.msg ( thehtml, 2 );
}
function wipeDoodle()
{
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);
}