We Keep Coding

Fill an Unknown Asymmetric Polygon with Pixel Manipulation (WebImage) in JS

Recently, I have been trying to create code to fill a polygon of any shape with color. I have gotten as far as being able to fill a shape that has lines of only one border size correctly, though I have found myself unable to do anything more than that. The problem is that the code does not know when to consider a line of pixels greater than that which it expects as a vertical or horizontal border of the shape. I am going through each pixel of the shape from left to right and checking if any of the pixels have any form of color by checking if the alpha value is 0 or not. Once it finds a pixel that does have an alpha value of anything other than 0, it moves forward a single pixel and then uses the even/odd technique to determine whether the point is inside part of the polygon or not (it makes an infinite line to the right and determines if the number of collisions with colored lines is odd, and if it is, the point is inside the polygon). In general, we consider a single, lone pixel to count as a single line, and we consider a horizontal line of more than one pixel to be two lines because of how often horizontal lines will be part of a border or not. Take the following scenario:
Here, the red dot is the point (pixel) we begin testing from. If we did not consider that horizontal line in the middle to be two points (as is shown by the red lines and x's), we would only have two points of intersection and therefore would not fill the pixel despite the fact that we most definitely do want to fill that pixel. As stated earlier, however, this brings up another problem with a different scenario:
In this case, if we do count a horizontal line of more than one pixel to be two separate lines, we end up not filling any areas with borders that are thicker than the expected thickness. For your reference, the function to handle this is as follows:
//imgData is essentially a WebImage object (explained more below) and r, g, and b are the color values for the fill color
function fillWithColor(imgData, r, g, b) {
//Boolean determining whether we should color the given pixel(s) or not
var doColor = false;
//Booleans determining whether the last pixel found in the entire image was colored
var blackLast = false;
//Booleans determining whether the last 1 or 2 pixels found after a given pixel were colored
var foundBlackPrev, foundBlackPrev2 = false;
//The number of colored pixels found
var blackCount = 0;
//Loop through the entire canvas
for(var y = 0; y < imgData.height; y += IMG_SCALE) {
for(var x = 0; x < imgData.width; x += IMG_SCALE) {
//Test if given pixel is colored
if(getAlpha(imgData, x, y) != 0) {
//If the last pixel was black, begin coloring
if(!blackLast) {
blackLast = true;
doColor = true;
}
} else {
//If the current pixel is not colored, but the last one was, find all colored lines to the right
if(blackLast){
for(var i = x; i < imgData.width; i += IMG_SCALE) {
//If the pixel is colored...
if(getAlpha(imgData, i, y) != 0) {
//If no colored pixel was found before, add to the count
if(!foundBlackPrev){
blackCount++;
foundBlackPrev = true;
} else {
//Otherwise, at least 2 colored pixels have been found in a row
foundBlackPrev2 = true;
}
} else {
//If two or more colored pixels were found in a row, add to the count
if(foundBlackPrev2) {
blackCount++;
}
//Reset the booleans
foundBlackPrev2 = foundBlackPrev = false;
}
}
}
//If the count is odd, we start coloring
if(blackCount & 1) {
blackCount = 0;
doColor = true;
} else {
//If the last pixel in the entire image was black, we stop coloring
if(blackLast) {
doColor = false;
}
}
//Reset the boolean
blackLast = false;
//If we are to be coloring the pixel, color it
if(doColor) {
//Color the pixel
for(var j = 0; j < IMG_SCALE; j++) {
for(var k = 0; k < IMG_SCALE; k++) {
//This is the same as calling setRed, setGreen, setBlue and setAlpha functions from the WebImage API all at once (parameters in order are WebImage object equivalent, x position of pixel, y position of pixel, red value, green value, blue value, and alpha value)
setRGB(imgData, x + j, y + k, r, g, b, 255);
}
}
}
}
}
}
//Update the image (essentially the same as removing all elements from the given area and calling add on the image)
clearCanvas();
putImageData(imgData, 0, 0, imgData.width, imgData.height);
//Return the modified data
return imgData;
}
Where...
imgData is the collection of all of the pixels in the given area (essentially a WebImage object)
IMG_SCALE is the integer value by which the image has been scaled up (which gives us the scale of the pixels as well). In this example, it is equal to 4 because the image is scaled up to 192x256 (from 48x64). This means that every "pixel" you see in the image is actually comprised of a 4x4 block of identically-colored pixels.
So, what I'm really looking for here is a way to determine whether a given colored pixel that comes after another is part of a horizontal border or if it is just another piece comprising the thickness of a vertical border. In addition, if I have the wrong approach to this problem in general, I would greatly appreciate any suggestions as to how to do this more efficiently. Thank you.

I understand the problem and I think you would do better if you would switch your strategy here. We know the following:
the point of start is inside the shape
the color should be filled for every pixel inside the shape
So, we could always push the neighbors of the current point into a queue to be processed and be careful to avoid processing the same points twice, this way traversing all the useful pixels and including them into the coloring plan. The function below is untested.
function fillColor(pattern, startingPoint, color, boundaryColor) {
let visitQueue = [];
let output = {};
if (startingPoint.x - 1 >= 0) visitQueue.push({startingPoint.x - 1, startingPoint.y});
if (startingPoint.x + 1 < pattern.width) visitQueue.push({startingPoint.x + 1, startingPoint.y});
if (startingPoint.y + 1 < pattern.height) visitQueue.push({startingPoint.x, startingPoint.y + 1});
if (startingPoint.y - 1 >= 0) visitQueue.push({startingPoint.x, startingPoint.y - 1});
let visited = {};
while (visitQueue.length > 0) {
let point = visitQueue[0];
visitQueue.shift();
if ((!visited[point.x]) || (visited[point.x].indexOf(point.y) < 0)) {
if (!visited[point.x]) visited[point.x] = [];
visited[point.x].push(point.y);
if (isBlank(pattern, point)) { //you need to implement isBlank
if (!output[point.x]) output[point.x] = [];
output[point.x].push(point.y);
if (point.x + 1 < pattern.width) visitQueue.push({point.x + 1, point.y});
if (point.x - 1 >= 0) visitQueue.push({point.x - 1, point.y});
if (point.y + 1 < pattern.height) visitQueue.push({point.x, point.y + 1});
if (point.y - 1 >= 0) visitQueue.push({point.x, point.y - 1})
}
}
}
return output;
}

As far as I understood you cannot "consider a horizontal line of more than one pixel to be two lines". I don't think you need to count black pixels the way you do, rather count groups of 1 or more pixels.
I would also tidy the code by avoiding using the "doColor" boolean variable. You could rather move the coloring code to a new function color(x,y) and call it straight away.

const ctx = document.querySelector("canvas").getContext("2d");
//ctx.lineWidth(10);//-as you asked we are setting greater border or line width,BUT "LINEWIDTH" IS NOT WORKING IN INBUILT STACKOVERFLOW SNIPPET USE IT IN A FILE I THINK STACKOVERFLOW IS NOT UP-TO-DATE,IN ANY IDE UNCOMENT THIS
ctx.beginPath();
ctx.moveTo(20, 20);
ctx.lineTo(250, 70);
ctx.lineTo(270, 120);
ctx.lineTo(170, 140);
ctx.lineTo(190, 80);
ctx.lineTo(100, 60);
ctx.lineTo(50, 130);
ctx.lineTo(20, 20);
ctx.stroke();
function getMousePosition(canvas, event) {
let rect = canvas.getBoundingClientRect();
let mx = event.clientX - rect.left;
let my = event.clientY - rect.top;
console.log("Coordinate x: " + mx, "Coordinate y: " + my);
floodFill(ctx, mx, my, [155, 0, 255, 255], 128);
}
let canvasElem = document.querySelector("canvas");
canvasElem.addEventListener("mousedown", function(e) {
getMousePosition(canvasElem, e);
});
function getPixel(imageData, x, y) {
if (x < 0 || y < 0 || x >= imageData.width || y >= imageData.height) {
return [-1, -1, -1, -1]; // impossible color
} else {
const offset = (y * imageData.width + x) * 4;
return imageData.data.slice(offset, offset + 4);
}
}
function setPixel(imageData, x, y, color) {
const offset = (y * imageData.width + x) * 4;
imageData.data[offset + 0] = color[0];
imageData.data[offset + 1] = color[1];
imageData.data[offset + 2] = color[2];
imageData.data[offset + 3] = color[0];
}
function colorsMatch(a, b, rangeSq) {
const dr = a[0] - b[0];
const dg = a[1] - b[1];
const db = a[2] - b[2];
const da = a[3] - b[3];
return dr * dr + dg * dg + db * db + da * da < rangeSq;
}
function floodFill(ctx, x, y, fillColor, range = 1) {
// read the pixels in the canvas
const imageData = ctx.getImageData(0, 0, ctx.canvas.width, ctx.canvas.height);
// flags for if we visited a pixel already
const visited = new Uint8Array(imageData.width, imageData.height);
// get the color we're filling
const targetColor = getPixel(imageData, x, y);
// check we are actually filling a different color
if (!colorsMatch(targetColor, fillColor)) {
const rangeSq = range * range;
const pixelsToCheck = [x, y];
while (pixelsToCheck.length > 0) {
const y = pixelsToCheck.pop();
const x = pixelsToCheck.pop();
const currentColor = getPixel(imageData, x, y);
if (!visited[y * imageData.width + x] &&
colorsMatch(currentColor, targetColor, rangeSq)) {
setPixel(imageData, x, y, fillColor);
visited[y * imageData.width + x] = 1; // mark we were here already
pixelsToCheck.push(x + 1, y);
pixelsToCheck.push(x - 1, y);
pixelsToCheck.push(x, y + 1);
pixelsToCheck.push(x, y - 1);
}
}
// put the data back
ctx.putImageData(imageData, 0, 0);
}
}
<canvas></canvas>
This is based on other answers
note:"LINEWIDTH" IS NOT WORKING IN INBUILT STACKOVERFLOW SNIPPET USE IT IN A FILE I THINK STACKOVERFLOW IS NOT UP-TO-DATE,
But it works well in simple HTML,JS website

JavaScript pixel by pixel canvas manipulation

I'm working on a simple web app which simplifies the colours of an uploaded image to a colour palette selected by the user. The script works, but it takes a really long time to loop through the whole image (for large images it's over a few minutes), changing the pixels.
Initially, I was writing to the canvas itself, but I changed the code so that changes are made to an ImageData object and the canvas is only updated at the end of the script. However, this didn't really make much difference.
// User selects colours:
colours = [[255,45,0], [37,36,32], [110,110,105], [18,96,4]];
function colourDiff(colour1, colour2) {
difference = 0
difference += Math.abs(colour1[0] - colour2[0]);
difference += Math.abs(colour1[1] - colour2[1]);
difference += Math.abs(colour1[2] - colour2[2]);
return(difference);
}
function getPixel(imgData, index) {
return(imgData.data.slice(index*4, index*4+4));
}
function setPixel(imgData, index, pixelData) {
imgData.data.set(pixelData, index*4);
}
data = ctx.getImageData(0,0,canvas.width,canvas.height);
for(i=0; i<(canvas.width*canvas.height); i++) {
pixel = getPixel(data, i);
lowestDiff = 1024;
lowestColour = [0,0,0];
for(colour in colours) {
colour = colours[colour];
difference = colourDiff(colour, pixel);
if(lowestDiff < difference) {
continue;
}
lowestDiff = difference;
lowestColour = colour;
}
console.log(i);
setPixel(data, i, lowestColour);
}
ctx.putImageData(data, 0, 0);
During the entire process, the website is completely frozen, so I can't even display a progress bar. Is there any way to optimise this so that it takes less time?

There is no need to slice the array each iteration. (As niklas has already stated).
I would loop over the data array instead of looping over the canvas dimensions and directly edit the array.
for(let i = 0; i < data.length; i+=4) { // i+=4 to step over each r,g,b,a pixel
let pixel = getPixel(data, i);
...
setPixel(data, i, lowestColour);
}
function setPixel(data, i, colour) {
data[i] = colour[0];
data[i+1] = colour[1];
data[i+2] = colour[2];
}
function getPixel(data, i) {
return [data[i], data[i+1], data[i+2]];
}
Also, console.log can bring a browser to it's knees if you've got the console open. If your image is 1920 x 1080 then you will be logging to the console 2,073,600 times.
You can also pass all of the processing off to a Web Worker for ultimate threaded performance. Eg. https://jsfiddle.net/pnmz75xa/

One problem or option for improvement is clearly your slice function, which will create a new array every time it is called, you do not need this. I would change the for loop like so:
for y in canvas.height {
for x in canvas.width {
//directly alter the canvas' pixels
}
}

Finding difference in color
I am adding an answer because you have use a very poor color match algorithm.
Finding how closely a color matches another is best done if you imagine each unique possible colour as a point in 3D space. The red, green, and blue values represent the x,y,z coordinate.
You can then use some basic geometry to locate the distance from one colour to the another.
// the two colours as bytes 0-255
const colorDist = (r1, g1, b1, r2, g2, b2) => Math.hypot(r1 - r2, g1 - g2, b1 - b2);
It is also important to note that the channel value 0-255 is a compressed value, the actual intensity is close to that value squared (channelValue ** 2.2). That means that red = 255 is 65025 times more intense than red = 1
The following function is a close approximation of the colour difference between two colors. Avoiding the Math.hypot function as it is very slow.
const pallet = [[1,2,3],[2,10,30]]; // Array of arrays representing rgb byte
// of the colors you are matching
function findClosest(r,g,b) {
var closest;
var dist = Infinity;
r *= r;
g *= g;
b *= b;
for (const col of pallet) {
const d = ((r - col[0] * col[0]) + (g - col[1] * col[1]) + (b - col[2] * col[2])) ** 0.5;
if (d < dist) {
if (d === 0) { // if same then return result
return col;
}
closest = col;
dist = d;
}
}
return closest;
}
As for performance, your best bet is either via a web worker, or use webGL to do the conversion in realtime.
If you want to keep it simple to prevent the code from blocking the page cut the job into smaller slices using a timer to allow the page breathing room.
The example uses setTimeout and performance.now() to do 10ms slices letting other page events and rendering to do there thing. It returns a promise that resolves when all pixels are processed
function convertBitmap(canvas, maxTime) { // maxTime in ms (1/1000 second)
return new Promise(allDone => {
const ctx = canvas.getContext("2d");
const pixels = ctx.getImageData(0, 0, canvas.width, canvas.height);
const data = pixels.data;
var idx = data.length / 4;
processPixels(); // start processing
function processPixels() {
const time = performance.now();
while (idx-- > 0) {
if (idx % 1024) { // check time every 1024 pixels
if (performance.now() - time > maxTime) {
setTimeout(processPixels, 0);
idx++;
return;
}
}
let i = idx * 4;
const col = findClosest(data[i], data[i + 1], data[i + 2]);
data[i++] = col[0];
data[i++] = col[1];
data[i] = col[2];
}
ctx.putImageData(pixels, 0, 0);
allDone("Pixels processed");
}
});
}
// process pixels in 10ms slices.
convertBitmap(myCanvas, 10).then(mess => console.log(mess));

HTML5 Canvas, better pixel control and better speed

I'm trying to create a little simulation with the help of HTML5 and Javascript using a canvas. My problem however is, I can't really think of a way to control the behavior of my pixels, without making every single pixel an object, which leads to an awful slowdown of my simulation.
Heres the code so far:
var pixels = [];
class Pixel{
constructor(color){
this.color=color;
}
}
window.onload=function(){
canv = document.getElementById("canv");
ctx = canv.getContext("2d");
createMap();
setInterval(game,1000/60);
};
function createMap(){
pixels=[];
for(i = 0; i <= 800; i++){
pixels.push(sub_pixels = []);
for(j = 0; j <= 800; j++){
pixels[i].push(new Pixel("green"));
}
}
pixels[400][400].color="red";
}
function game(){
ctx.fillStyle = "white";
ctx.fillRect(0,0,canv.width,canv.height);
for(i = 0; i <= 800; i++){
for(j = 0; j <= 800; j++){
ctx.fillStyle=pixels[i][j].color;
ctx.fillRect(i,j,1,1);
}
}
for(i = 0; i <= 800; i++){
for(j = 0; j <= 800; j++){
if(pixels[i][j].color == "red"){
direction = Math.floor((Math.random() * 4) + 1);
switch(direction){
case 1:
pixels[i][j-1].color= "red";
break;
case 2:
pixels[i+1][j].color= "red";
break;
case 3:
pixels[i][j+1].color= "red";
break;
case 4:
pixels[i-1][j].color= "red";
break;
}
}
}
}
}
function retPos(){
return Math.floor((Math.random() * 800) + 1);
}
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<script language="javascript" type="text/javascript" src="game.js"></script>
</head>
<body>
<canvas width="800px" height="800px" id="canv"></canvas>
</body>
</html>
So my two big questions are, what better way of controlling those pixels is there? And how can I speed up the pixel generation?
Hope you can help me.

Optimizing pixel manipulation
There are many options to speed up your code
Pixels as 32bit ints
The following will slug most machines with too much work.
// I removed fixed 800 and replaced with const size
for(i = 0; i <= size; i++){
for(j = 0; j <= size; j++){
ctx.fillStyle=pixels[i][j].color;
ctx.fillRect(i,j,1,1);
}
}
Don't write each pixel via a rect. Use the pixel data you can get from the canvas API via createImageData and associated functions. It uses typed arrays that are a little quicker than arrays and can have multiple view on the same content.
You can write all the pixels to the canvas in a single call. Not blindingly fast but a zillion times faster than what you are doing.
const size = 800;
const imageData = ctx.createImageData(size,size);
// get a 32 bit view
const data32 = new Uint32Array(imageData.data.buffer);
// To set a single pixel
data32[x+y*size] = 0xFF0000FF; // set pixel to red
// to set all pixels
data32.fill(0xFF00FF00); // set all to green
To get a pixel at a pixel coord
const pixel = data32[x + y * imageData.width];
See Accessing pixel data for more on using the image data.
The pixel data is not displayed until you put it onto the canvas
ctx.putImageData(imageData,0,0);
That will give you a major improvement.
Better data organization.
When performance is critical you sacrifice memory and simplicity to get more CPU cycles doing what you want and less doing a lot of nothing.
You have red pixels randomly expanding into the scene, you read every pixel and check (via a slow string comparison) if it is red. When you find one you add a random red pixel besides it.
Checking the green pixels is a waste and can be avoided. Expanding red pixels that are completely surrounded by other reds is also pointless. They do nothing.
The only pixels you are interested in are the red pixels that are next to green pixels.
Thus you can create a buffer that holds the location of all active red pixels, An active red has at least one green. Each frame you check all the active reds, spawning new ones if they can, and killing them if they are surrounded in red.
We don't need to store the x,y coordinate of each red, just the memory address so we can use a flat array.
const reds = new Uint32Array(size * size); // max size way over kill but you may need it some time.
You dont want to have to search for reds in your reds array so you need to keep track of how many active reds there are. You want all the active reds to be at the bottom of the array. You need to check each active red only once per frame. If a red is dead than all above it must move down one array index. But you only want to move each red only once per frame.
Bubble array
I dont know what this type of array is called its like a separation tank, dead stuff slowly moves up and live stuff moves down. Or unused items bubble up used items settle to the bottom.
I will show it as functional because it will be easier to understand. but is better implemented as one brute force function
// data32 is the pixel data
const size = 800; // width and height
const red = 0xFF0000FF; // value of a red pixel
const green = 0xFF00FF00; // value of a green pixel
const reds = new Uint32Array(size * size); // max size way over kill but you var count = 0; // total active reds
var head = 0; // index of current red we are processing
var tail = 0; // after a red has been process it is move to the tail
var arrayOfSpawnS = [] // for each neighbor that is green you want
// to select randomly to spawn to. You dont want
// to spend time processing so this is a lookup
// that has all the possible neighbor combinations
for(let i = 0; i < 16; i ++){
let j = 0;
const combo = [];
i & 1 && (combo[j++] = 1); // right
i & 2 && (combo[j++] = -1); // left
i & 4 && (combo[j++] = -size); // top
i & 5 && (combo[j++] = size); // bottom
arrayOfSpawnS.push(combo);
}
function addARed(x,y){ // add a new red
const pixelIndex = x + y * size;
if(data32[pixelIndex] === green) { // check if the red can go there
reds[count++] = pixelIndex; // add the red with the pixel index
data32[pixelIndex] = red; // and set the pixel
}
}
function safeAddRed(pixelIndex) { // you know that some reds are safe at the new pos so a little bit faster
reds[count++] = pixelIndex; // add the red with the pixel index
data32[pixelIndex] = red; // and set the pixel
}
// a frame in the life of a red. Returns false if red is dead
function processARed(indexOfRed) {
// get the pixel index
var pixelIndex = reds[indexOfRed];
// check reds neighbors right left top and bottom
// we fill a bit value with each bit on if there is a green
var n = data32[pixelIndex + 1] === green ? 1 : 0;
n += data32[pixelIndex - 1] === green ? 2 : 0;
n += data32[pixelIndex - size] === green ? 4 : 0;
n += data32[pixelIndex + size] === green ? 8 : 0;
if(n === 0){ // no room to spawn so die
return false;
}
// has room to spawn so pick a random
var nCount = arrayOfSpawnS[n].length;
// if only one spawn point then rather than spawn we move
// this red to the new pos.
if(nCount === 1){
reds[indexOfRed] += arrayOfSpawnS[n][0]; // move to next pos
}else{ // there are several spawn points
safeAddRed(pixelIndex + arrayOfSpawnS[n][(Math.random() * nCount)|0]);
}
// reds frame is done so return still alive to spawn another frame
return true;
}
Now to process all the reds.
This is the heart of the bubble array. head is used to index each active red. tail is the index of where to move the current head if no deaths have been encountered tail is equal to head. If however a dead item is encountered the head move up one while the tail remains pointing to the dead item. This moves all the active items to the bottom.
When head === count all active items have been checked. The value of tail now contains the new count which is set after the iteration.
If you were using an object rather than a Integer, instead of moving the active item down you swap the head and tail items. This effectively creates a pool of available objects that can be used when adding new items. This type of array management incurs not GC or Allocation overhead and is hence very quick when compared to stacks and object pools.
function doAllReds(){
head = tail = 0; // start at the bottom
while(head < count){
if(processARed(head)){ // is red not dead
reds[tail++] = reds[head++]; // move red down to the tail
}else{ // red is dead so this creates a gap in the array
// Move the head up but dont move the tail,
// The tail is only for alive reds
head++;
}
}
// All reads done. The tail is now the new count
count = tail;
}
The Demo.
The demo will show you the speed improvement. I used the functional version and there could be some other tweaks.
You can also consider webWorkers to get event more speed. Web worker run on a separate javascript context and provides true concurrent processing.
For the ultimate speed use WebGL. All the logic can be done via a fragment shader on the GPU. This type of task is very well suited to parallel processing for which the GPU is designed.
Will be back later to clean up this answer (got a little too long)
I have also added a boundary to the pixel array as the reds were spawning off the pixel array.
const size = canvas.width;
canvas.height = canvas.width;
const ctx = canvas.getContext("2d");
const red = 0xFF0000FF;
const green = 0xFF00FF00;
const reds = new Uint32Array(size * size);
const wall = 0xFF000000;
var count = 0;
var head = 0;
var tail = 0;
var arrayOfSpawnS = []
for(let i = 0; i < 16; i ++){
let j = 0;
const combo = [];
i & 1 && (combo[j++] = 1); // right
i & 2 && (combo[j++] = -1); // left
i & 4 && (combo[j++] = -size); // top
i & 5 && (combo[j++] = size); // bottom
arrayOfSpawnS.push(combo);
}
const imageData = ctx.createImageData(size,size);
const data32 = new Uint32Array(imageData.data.buffer);
function createWall(){//need to keep the reds walled up so they dont run free
for(let j = 0; j < size; j ++){
data32[j] = wall;
data32[j * size] = wall;
data32[j * size + size - 1] = wall;
data32[size * (size - 1) +j] = wall;
}
}
function addARed(x,y){
const pixelIndex = x + y * size;
if (data32[pixelIndex] === green) {
reds[count++] = pixelIndex;
data32[pixelIndex] = red;
}
}
function processARed(indexOfRed) {
var pixelIndex = reds[indexOfRed];
var n = data32[pixelIndex + 1] === green ? 1 : 0;
n += data32[pixelIndex - 1] === green ? 2 : 0;
n += data32[pixelIndex - size] === green ? 4 : 0;
n += data32[pixelIndex + size] === green ? 8 : 0;
if(n === 0) { return false }
var nCount = arrayOfSpawnS[n].length;
if (nCount === 1) { reds[indexOfRed] += arrayOfSpawnS[n][0] }
else {
pixelIndex += arrayOfSpawnS[n][(Math.random() * nCount)|0]
reds[count++] = pixelIndex;
data32[pixelIndex] = red;
}
return true;
}
function doAllReds(){
head = tail = 0;
while(head < count) {
if(processARed(head)) { reds[tail++] = reds[head++] }
else { head++ }
}
count = tail;
}
function start(){
data32.fill(green);
createWall();
var startRedCount = (Math.random() * 5 + 1) | 0;
for(let i = 0; i < startRedCount; i ++) { addARed((Math.random() * size-2+1) | 0, (Math.random() * size-2+1) | 0) }
ctx.putImageData(imageData,0,0);
setTimeout(doItTillAllDead,1000);
countSameCount = 0;
}
var countSameCount;
var lastCount;
function doItTillAllDead(){
doAllReds();
ctx.putImageData(imageData,0,0);
if(count === 0 || countSameCount === 100){ // all dead
setTimeout(start,1000);
}else{
countSameCount += count === lastCount ? 1 : 0;
lastCount = count; //
requestAnimationFrame(doItTillAllDead);
}
}
start();
<canvas width="800" height="800" id="canvas"></canvas>

The main cause of your slow down is your assumption that you need to loop over every pixel for every operation. You do not do this, as that would be 640,000 iterations for every operation you need to do.
You also shouldn't be doing any manipulation logic within the render loop. The only thing that should be there is drawing code. So this should be moved out to preferably a separate thread (Web Workers). If unable to use those a setTimeout/Interval call.
So first a couple of small changes:
Make Pixel class contain the pixel's coordinates along with the color:
class Pixel{
constructor(color,x,y){
this.color=color;
this.x = x;
this.y = y;
}
}
Keep an array of pixels that will end up creating new red pixels. And another one to keep track of what pixels have been updated so we know which ones need drawn.
var pixels = [];
var infectedPixesl = [];
var updatedPixels = [];
Now the easiest part of the code to change is the render loop. Since the only thing that it needs to do is draw the pixels it will be only a couple lines.
function render(){
var numUpdatedPixels = updatedPixels.length;
for(let i=0; i<numUpdatedPixels; i++){
let pixel = updatedPixels[i];
ctx.fillStyle = pixel.color;
ctx.fillRect(pixel.x,pixel.y,1,1);
}
//clear out the updatedPixels as they should no longer be considered updated.
updatedPixels = [];
//better method than setTimeout/Interval for drawing
requestAnimationFrame(render);
}
From there we can move on to the logic. We will loop over the infectedPixels array, and with each pixel we decide a random direction and get that pixel. If this selected pixel is red we do nothing and continue on. Otherwise we change it's color and add it to a temporary array affectedPixels. After which we test to see if all the pixels around the original pixel are all red, if so we can remove it from the infectedPixels as there is no need to check it again. Then add all the pixels from affectedPixels onto the infectedPixels as these are now new pixels that need to be checked. And the last step is to also add affectedPixels onto updatedPixels so that the render loop draws the changes.
function update(){
var affectedPixels = [];
//needed as we shouldn't change an array while looping over it
var stillInfectedPixels = [];
var numInfected = infectedPixels.length;
for(let i=0; i<numInfected; i++){
let pixel = infectedPixels[i];
let x = pixel.x;
let y = pixel.y;
//instead of using a switch statement, use the random number as the index
//into a surroundingPixels array
let surroundingPixels = [
(pixels[x] ? pixels[x][y - 1] : null),
(pixels[x + 1] ? pixels[x + 1][y] : null),
(pixels[x] ? pixels[x][y + 1] : null),
(pixels[x - 1] ? pixels[x - 1][y] : null)
].filter(p => p);
//filter used above to remove nulls, in the cases of edge pixels
var rand = Math.floor((Math.random() * surroundingPixels.length));
let selectedPixel = surroundingPixels[rand];
if(selectedPixel.color == "green"){
selectedPixel.color = "red";
affectedPixels.push(selectedPixel);
}
if(!surroundingPixels.every(p=>p.color=="red")){
stillInfectedPixels.push(pixel);
}
}
infectedPixels = stillInfectedPixel.concat( affectedPixels );
updatedPixels.push(...affectedPixels);
}
Demo
var pixels = [],
infectedPixels = [],
updatedPixels = [],
canv, ctx;
window.onload = function() {
canv = document.getElementById("canv");
ctx = canv.getContext("2d");
createMap();
render();
setInterval(() => {
update();
}, 16);
};
function createMap() {
for (let y = 0; y < 800; y++) {
pixels.push([]);
for (x = 0; x < 800; x++) {
pixels[y].push(new Pixel("green",x,y));
}
}
pixels[400][400].color = "red";
updatedPixels = [].concat(...pixels);
infectedPixels.push(pixels[400][400]);
}
class Pixel {
constructor(color, x, y) {
this.color = color;
this.x = x;
this.y = y;
}
}
function update() {
var affectedPixels = [];
var stillInfectedPixels = [];
var numInfected = infectedPixels.length;
for (let i = 0; i < numInfected; i++) {
let pixel = infectedPixels[i];
let x = pixel.x;
let y = pixel.y;
let surroundingPixels = [
(pixels[x] ? pixels[x][y - 1] : null),
(pixels[x + 1] ? pixels[x + 1][y] : null),
(pixels[x] ? pixels[x][y + 1] : null),
(pixels[x - 1] ? pixels[x - 1][y] : null)
].filter(p => p);
var rand = Math.floor((Math.random() * surroundingPixels.length));
let selectedPixel = surroundingPixels[rand];
if (selectedPixel.color == "green") {
selectedPixel.color = "red";
affectedPixels.push(selectedPixel);
}
if (!surroundingPixels.every(p => p.color == "red")) {
stillInfectedPixels.push(pixel);
}
}
infectedPixels = stillInfectedPixels.concat(affectedPixels);
updatedPixels.push(...affectedPixels);
}
function render() {
var numUpdatedPixels = updatedPixels.length;
for (let i = 0; i < numUpdatedPixels; i++) {
let pixel = updatedPixels[i];
ctx.fillStyle = pixel.color;
ctx.fillRect(pixel.x, pixel.y, 1, 1);
}
updatedPixels = [];
requestAnimationFrame(render);
}
<canvas id="canv" width="800" height="800"></canvas>

Find regions of similar color in image

I've been working on this problem for some time now with little promising results. I am trying to split up an image into connected regions of similar color. (basically split a list of all the pixels into multiple groups (each group containing the coordinates of the pixels that belong to it and share a similar color).
For example:
http://unsplash.com/photos/SoC1ex6sI4w/
In this image the dark clouds at the top would probably fall into one group. Some of the grey rock on the mountain in another, and some of the orange grass in another. The snow would be another - the red of the backpack - etc.
I'm trying to design an algorithm that will be both accurate and efficient (it needs to run in a matter of ms on midrange laptop grade hardware)
Below is what I have tried:
Using a connected component based algorithm to go through every pixel from top left scanning every line of pixels from left to right (and comparing the current pixel to the top pixel and left pixel). Using the CIEDE2000 color difference formula if the pixel at the top or left was within a certain range then it would be considered "similar" and part of the group.
This sort of worked - but the problem is it relies on color regions having sharp edges - if any color groups are connected by a soft gradient it will travel down that gradient and continue to "join" the pixels as the difference between the individual pixels being compared is small enough to be considered "similar".
To try to fix this I chose to set every visited pixel's color to the color of most "similar" adjacent pixel (either top or left). If there are no similar pixels than it retains it's original color. This somewhat fixes the issue of more blurred boundaries or soft edges because the first color of a new group will be "carried" along as the algorithm progresses and eventually the difference between that color and the current compared color will exceed the "similarity" threashold and no longer be part of that group.
Hopefully this is making sense. The problem is neither of these options are really working. On the image above what is returned are not clean groups but noisy fragmented groups that is not what I am looking for.
I'm not looking for code specifically - but more ideas as to how an algorithm could be structured to successfully combat this problem. Does anyone have ideas about this?
Thanks!

You could convert from RGB to HSL to make it easier to calculate the distance between the colors. I'm setting the color difference tolerance in the line:
if (color_distance(original_pixels[i], group_headers[j]) < 0.3) {...}
If you change 0.3, you can get different results.
See it working.
Please, let me know if it helps.
function hsl_to_rgb(h, s, l) {
// from http://stackoverflow.com/questions/2353211/hsl-to-rgb-color-conversion
var r, g, b;
if (s == 0) {
r = g = b = l; // achromatic
} else {
var hue2rgb = function hue2rgb(p, q, t) {
if (t < 0) t += 1;
if (t > 1) t -= 1;
if (t < 1 / 6) return p + (q - p) * 6 * t;
if (t < 1 / 2) return q;
if (t < 2 / 3) return p + (q - p) * (2 / 3 - t) * 6;
return p;
}
var q = l < 0.5 ? l * (1 + s) : l + s - l * s;
var p = 2 * l - q;
r = hue2rgb(p, q, h + 1 / 3);
g = hue2rgb(p, q, h);
b = hue2rgb(p, q, h - 1 / 3);
}
return [Math.round(r * 255), Math.round(g * 255), Math.round(b * 255)];
}
function rgb_to_hsl(r, g, b) {
// from http://stackoverflow.com/questions/2353211/hsl-to-rgb-color-conversion
r /= 255, g /= 255, b /= 255;
var max = Math.max(r, g, b),
min = Math.min(r, g, b);
var h, s, l = (max + min) / 2;
if (max == min) {
h = s = 0; // achromatic
} else {
var d = max - min;
s = l > 0.5 ? d / (2 - max - min) : d / (max + min);
switch (max) {
case r:
h = (g - b) / d + (g < b ? 6 : 0);
break;
case g:
h = (b - r) / d + 2;
break;
case b:
h = (r - g) / d + 4;
break;
}
h /= 6;
}
return [h, s, l];
}
function color_distance(v1, v2) {
// from http://stackoverflow.com/a/13587077/1204332
var i,
d = 0;
for (i = 0; i < v1.length; i++) {
d += (v1[i] - v2[i]) * (v1[i] - v2[i]);
}
return Math.sqrt(d);
};
function round_to_groups(group_nr, x) {
var divisor = 255 / group_nr;
return Math.ceil(x / divisor) * divisor;
};
function pixel_data_to_key(pixel_data) {
return pixel_data[0].toString() + '-' + pixel_data[1].toString() + '-' + pixel_data[2].toString();
}
function posterize(context, image_data, palette) {
for (var i = 0; i < image_data.data.length; i += 4) {
rgb = image_data.data.slice(i, i + 3);
hsl = rgb_to_hsl(rgb[0], rgb[1], rgb[2]);
key = pixel_data_to_key(hsl);
if (key in palette) {
new_hsl = palette[key];
new_rgb = hsl_to_rgb(new_hsl[0], new_hsl[1], new_hsl[2]);
rgb = hsl_to_rgb(hsl);
image_data.data[i] = new_rgb[0];
image_data.data[i + 1] = new_rgb[1];
image_data.data[i + 2] = new_rgb[2];
}
}
context.putImageData(image_data, 0, 0);
}
function draw(img) {
var canvas = document.getElementById('canvas');
var context = canvas.getContext('2d');
context.drawImage(img, 0, 0, canvas.width, canvas.height);
img.style.display = 'none';
var image_data = context.getImageData(0, 0, canvas.width, canvas.height);
var data = image_data.data;
context.drawImage(target_image, 0, 0, canvas.width, canvas.height);
data = context.getImageData(0, 0, canvas.width, canvas.height).data;
original_pixels = [];
for (i = 0; i < data.length; i += 4) {
rgb = data.slice(i, i + 3);
hsl = rgb_to_hsl(rgb[0], rgb[1], rgb[2]);
original_pixels.push(hsl);
}
group_headers = [];
groups = {};
for (i = 0; i < original_pixels.length; i += 1) {
if (group_headers.length == 0) {
group_headers.push(original_pixels[i]);
}
group_found = false;
for (j = 0; j < group_headers.length; j += 1) {
// if a similar color was already observed
if (color_distance(original_pixels[i], group_headers[j]) < 0.3) {
group_found = true;
if (!(pixel_data_to_key(original_pixels[i]) in groups)) {
groups[pixel_data_to_key(original_pixels[i])] = group_headers[j];
}
}
if (group_found) {
break;
}
}
if (!group_found) {
if (group_headers.indexOf(original_pixels[i]) == -1) {
group_headers.push(original_pixels[i]);
}
if (!(pixel_data_to_key(original_pixels[i]) in groups)) {
groups[pixel_data_to_key(original_pixels[i])] = original_pixels[i];
}
}
}
posterize(context, image_data, groups)
}
var target_image = new Image();
target_image.crossOrigin = "";
target_image.onload = function() {
draw(target_image)
};
target_image.src = "http://i.imgur.com/zRzdADA.jpg";
canvas {
width: 300px;
height: 200px;
}
<canvas id="canvas"></canvas>

You can use "Mean Shift Filtering" algorithm to do the same.
Here's an example.
You will have to determine function parameters heuristically.
And here's the wrapper for the same in node.js
npm Wrapper for meanshift algorithm
Hope this helps!

The process you are trying to complete is called Image Segmentation and it's a well studied area in computer vision, with hundreds of different algorithms and implementations.
The algorithm you mentioned should work for simple images, however for real world images such as the one you linked to, you will probably need a more sophisticated algorithm, maybe even one that is domain specific (are all of your images contains a view?).
I have little experience in Node.js, however from Googling a bit I found the GraphicsMagic library, which as a segment function that might do the job (haven't verified).
In any case, I would try looking for "Image segmentation" libraries, and if possible, not limit myself only to Node.js implementations, as this language is not the common practice for writing vision applications, as opposed to C++ / Java / Python.

I would try a different aproach. Check out this description of how a flood fill algorithm could work:
Create an array to hold information about already colored coordinates.
Create a work list array to hold coordinates that must be looked at. Put the start position in it.
When the work list is empty, we are done.
Remove one pair of coordinates from the work list.
If those coordinates are already in our array of colored pixels, go back to step 3.
Color the pixel at the current coordinates and add the coordinates to the array of colored pixels.
Add the coordinates of each adjacent pixel whose color is the same as the starting pixel’s original color to the work list.
Return to step 3.
The "search approach" is superior because it does not only search from left to right, but in all directions.

You might look at k-means clustering.
http://docs.opencv.org/3.0-beta/modules/core/doc/clustering.html

algorithm to randomly & efficiently place 100 circles without any overlap?

I am trying to write a script to place 100 circles of varying sizes onto a stage. I've outlined the concise requirements below.
Given the following:
var stage; // contains a "width" and "height" property.
var circle; // the circle class. contains x, y, radius & a unique id property.
var circleArray; // contains 100 circle instances
requirements:
write a function to place 100 circles of varying radius onto the stage.
placements must be random but evenly distributed (no clumping).
placement must be performant - this will be executing on a mobile web browser.
circles must not intersect/overlap other circles.
circle.x >= 0 must be true.
circle.y >= 0 && circle.y <= stage.height must be true.
circles may have any of the following radius sizes (assigned at creation):
150
120
90
80
65
My current attempt is a brute-force method, which does not operate efficiently. If I attempt to insert any more than ~10 circles, the browser hangs. Below is my current implementation, which I am completely OK with throwing away in favor of a more performant / better one.
Here is a live demo (NOTE: there is no actual drawing code, just the logic, but it will still lock up the browser so be warned!!) http://jsbin.com/muhiziduxu/2/edit?js,console
function adjustForOverlap (circleArray) {
// a reference to the circle that is invoking this function.
var _this = this;
// remove this circle from the array we are iterating over.
var arr = circleArray.filter(function (circle){
return circle.id !== _this.id;
});
// while repeat == true, the circle may be overlapping something.
var repeat = true;
while(repeat) {
var hasOverlap = false;
for (var i=0; i<arr.length; i++) {
var other = arr[i];
var dx = _self.x - other.x;
var dy = _self.y - other.y;
var rr = _self.radius + other.radius;
if (dx * dx + dy * dy < rr * rr) {
// if here, then an overlap was detected.
hit = true;
break;
}
}
// if hit is false, the circle didn't overlap anything, so break.
if (hit === false) {
repeat = false;
break;
} else {
// an overlap was detected, so randomize position.
_self.x = Math.random() * (stage.width*2);
_self.y = Math.random() * stage.height;
}
}
}

There are lots of efficient collision detection algorithms. Many of them work by dividing up the space into cells and maintaining a separate data structure with efficient lookup of other objects in the cell. The basic steps are:
Identify a random spot for your new circle
Determine which cells it's in
Look in each of those cells for a collision
If there's a collision, goto 1.
Else, add the new circle to each of the cells it overlaps.
You can use a simple square grid (i.e. a 2-d array) for the cell data structure, or something else like a quadtree. You can also in some cases get a bit of extra speed by trying a cheap-but-coarse collision check first (do the bounding boxes overlap), and if that returns true try the slightly more expensive and exact check.
Update
For quadtrees, check out d3-quadtree, which ought to give you a pretty good implementation, with examples.
For a (very quick, untested) 2-d array implementation:
function Grid(radius, width, height) {
// I'm not sure offhand how to find the optimum grid size.
// Let's use a radius as a starting point
this.gridX = Math.ceil(width / radius);
this.gridY = Math.ceil(height / radius);
// Determine cell size
this.cellWidth = width / this.gridX;
this.cellHeight = height / this.gridY;
// Create the grid structure
this.grid = [];
for (var i = 0; i < gridY; i++) {
// grid row
this.grid[i] = [];
for (var j = 0; j < gridX; j++) {
// Grid cell, holds refs to all circles
this.grid[i][j] = [];
}
}
}
Grid.prototype = {
// Return all cells the circle intersects. Each cell is an array
getCells: function(circle) {
var cells = [];
var grid = this.grid;
// For simplicity, just intersect the bounding boxes
var gridX1Index = Math.floor(
(circle.x - circle.radius) / this.cellWidth
);
var gridX2Index = Math.ceil(
(circle.x + circle.radius) / this.cellWidth
);
var gridY1Index = Math.floor(
(circle.y - circle.radius) / this.cellHeight
);
var gridY2Index = Math.ceil(
(circle.y + circle.radius) / this.cellHeight
);
for (var i = gridY1Index; i < gridY2Index; i++) {
for (var j = gridX1Index; j < gridX2Index; j++) {
// Add cell to list
cells.push(grid[i][j]);
}
}
return cells;
},
add: function(circle) {
this.getCells(circle).forEach(function(cell) {
cell.push(circle);
});
},
hasCollisions: function(circle) {
return this.getCells(circle).some(function(cell) {
return cell.some(function(other) {
return this.collides(circle, other);
}, this);
}, this);
},
collides: function (circle, other) {
if (circle === other) {
return false;
}
var dx = circle.x - other.x;
var dy = circle.y - other.y;
var rr = circle.radius + other.radius;
return (dx * dx + dy * dy < rr * rr);
}
};
var g = new Grid(150, 1000, 800);
g.add({x: 100, y: 100, radius: 50});
g.hasCollisions({x: 100, y:80, radius: 100});
Here's a fully-functional example: http://jsbin.com/cojoxoxufu/1/edit?js,output
Note that this only shows 30 circles. It looks like the problem is often unsolvable with your current radii, width, and height. This is set up to look for up to 500 locations for each circle before giving up and accepting a collision.