关于图像处理:JavaScript中的单色抖动(Bayer,Atkinson,Floyd–Steinberg)

monochrome dithering in JavaScript (Bayer, Atkinson, Floyd–Steinberg)

我正在使用HTML5中的摄像头过滤器。对于老式Mac感觉,Atkinson抖动效果很好。

meemoo

我仔细阅读了算法,但在将伪代码转换为JavaScript时遇到了麻烦:

1
2
3
4
5
for each y
  for each x
    oldpixel := pixel[x][y] + threshold_map_4x4[x mod 4][y mod 4]
    newpixel := find_closest_palette_color(oldpixel)
    pixel[x][y] := newpixel

在AS3,PHP或JS中是否有示例?您能否解释threshold_map_4x4[x mod 4][y mod 4]发生了什么?

gameboy(由Meemoo Gameboy GIFerizer制造)

弄清楚了。在Wikipedia中,它表示"例如,在单色渲染中,如果像素的值(缩放到0-9范围内)小于矩阵相应单元格中的像素数,则将该像素绘制为黑色,否则将其绘制为黑色白色。"在js中,通过平均当前像素(0-255)和地图值(15-240)并将其与阈值(通常为129)进行比较,我获得了良好的效果:

1
2
var map = (imageData.data[currentPixel] + bayerThresholdMap[x%4][y%4]) / 2;
imageData.data[currentPixel] = (map < threshold) ? 0 : 255;

此处是我使用不同算法的整个单色函数:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
var bayerThresholdMap = [
  [  15, 135,  45, 165 ],
  [ 195,  75, 225, 105 ],
  [  60, 180,  30, 150 ],
  [ 240, 120, 210,  90 ]
];

var lumR = [];
var lumG = [];
var lumB = [];
for (var i=0; i<256; i++) {
  lumR[i] = i*0.299;
  lumG[i] = i*0.587;
  lumB[i] = i*0.114;
}

function monochrome(imageData, threshold, type){

  var imageDataLength = imageData.data.length;

  // Greyscale luminance (sets r pixels to luminance of rgb)
  for (var i = 0; i <= imageDataLength; i += 4) {
    imageData.data[i] = Math.floor(lumR[imageData.data[i]] + lumG[imageData.data[i+1]] + lumB[imageData.data[i+2]]);
  }

  var w = imageData.width;
  var newPixel, err;

  for (var currentPixel = 0; currentPixel <= imageDataLength; currentPixel+=4) {

    if (type ==="none") {
      // No dithering
      imageData.data[currentPixel] = imageData.data[currentPixel] < threshold ? 0 : 255;
    } else if (type ==="bayer") {
      // 4x4 Bayer ordered dithering algorithm
      var x = currentPixel/4 % w;
      var y = Math.floor(currentPixel/4 / w);
      var map = Math.floor( (imageData.data[currentPixel] + bayerThresholdMap[x%4][y%4]) / 2 );
      imageData.data[currentPixel] = (map < threshold) ? 0 : 255;
    } else if (type ==="floydsteinberg") {
      // Floyda€"Steinberg dithering algorithm
      newPixel = imageData.data[currentPixel] < 129 ? 0 : 255;
      err = Math.floor((imageData.data[currentPixel] - newPixel) / 16);
      imageData.data[currentPixel] = newPixel;

      imageData.data[currentPixel       + 4 ] += err*7;
      imageData.data[currentPixel + 4*w - 4 ] += err*3;
      imageData.data[currentPixel + 4*w     ] += err*5;
      imageData.data[currentPixel + 4*w + 4 ] += err*1;
    } else {
      // Bill Atkinson's dithering algorithm
      newPixel = imageData.data[currentPixel] < threshold ? 0 : 255;
      err = Math.floor((imageData.data[currentPixel] - newPixel) / 8);
      imageData.data[currentPixel] = newPixel;

      imageData.data[currentPixel       + 4 ] += err;
      imageData.data[currentPixel       + 8 ] += err;
      imageData.data[currentPixel + 4*w - 4 ] += err;
      imageData.data[currentPixel + 4*w     ] += err;
      imageData.data[currentPixel + 4*w + 4 ] += err;
      imageData.data[currentPixel + 8*w     ] += err;
    }

    // Set g and b pixels equal to r
    imageData.data[currentPixel + 1] = imageData.data[currentPixel + 2] = imageData.data[currentPixel];
  }

  return imageData;
}

我很欣赏优化提示。


这是我所有的单色抖动功能,可用作网络工作者:https://github.com/meemoo/meemooapp/blob/master/src/nodes/image-monochrome-worker.js

使用网络摄像头进行实时演示:http://meemoo.org/iframework/#gist/3721129


我将其作为调试代码进行操作:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
var canvas = document.createElement('canvas');
var ctx    = canvas.getContext('2d');

var img    = new Image();
img.src    ="http://i.stack.imgur.com/tHDY8.png";
img.onload = function() {
    canvas.width  = this.width;
    canvas.height = this.height;
    ctx.drawImage( this, 0, 0, this.width, this.height );

    var imageData  = ctx.getImageData( 0, 0, this.width, this.height);
    var depth      = 32;


    // Matrix
    var threshold_map_4x4 = [
        [  1,  9,  3, 11 ],
        [ 13,  5, 15,  7 ],
        [  4, 12,  2, 10 ],
        [ 16,  8, 14,  6 ]
    ];

    // imageData
    var width  = imageData.width;
    var height = imageData.height;
    var pixel  = imageData.data;
    var x, y, a, b;

    // filter
    for ( x=0; x<width; x++ )
    {
        for ( y=0; y<height; y++ )
        {
            a    = ( x * height + y ) * 4;
            b    = threshold_map_4x4[ x%4 ][ y%4 ];
            pixel[ a + 0 ] = ( (pixel[ a + 0 ]+ b) / depth | 0 ) * depth;
            pixel[ a + 1 ] = ( (pixel[ a + 1 ]+ b) / depth | 0 ) * depth;
            pixel[ a + 2 ] = ( (pixel[ a + 2 ]+ b) / depth | 0 ) * depth;
            //pixel[ a + 3 ] = ( (pixel[ a + 3 ]+ b) / depth | 3 ) * depth;
        }
    }

    ctx.putImageData( imageData, 0, 0);

};

document.body.appendChild(canvas);

似乎工作正常,您可以更改depth变量以更改后代化。