https://github.com/nathsou/WebCV

WebCV - Blazingly fast computer vision in the browser
https://github.com/nathsou/WebCV

Last synced: about 2 months ago
JSON representation

WebCV - Blazingly fast computer vision in the browser

• Host: GitHub
• URL: https://github.com/nathsou/WebCV
• Owner: nathsou
• Created: 2016-03-05T21:48:48.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2016-03-06T01:24:45.000Z (over 8 years ago)
• Last Synced: 2024-05-05T05:32:42.315Z (5 months ago)
• Language: JavaScript
• Homepage:
• Size: 1.69 MB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-blazingly-fast - WebCV - WebCV - Blazingly fast computer vision in the browser (JavaScript)

README

        
## WebCV - Blazingly fast computer vision in the browser

WebCV is a [computer vision](https://en.wikipedia.org/wiki/Computer_vision) library written in [TypeScript](http://www.typescriptlang.org/)

and compiled into JavaScript, it uses the [WebGL](https://developer.mozilla.org/en/docs/Web/API/WebGL_API) API and it's GLSL ES 2.0 capacities

in order to apply filters to images using the GPU.

### The basics

##### Creating a *Context* :

The *Context* class initiates some default GLSL vertex and fragment shaders required for any GLSL program.

In order to create a *Context*, one needs to provide WebCV with an HTML5 canvas element and at least one image :

```javascript

var tex = new wcv.Texture('#lena'),

	ctx = new wcv.Context('#canvas', tex);

	

var fragmentShader = ctx.fs,

	vertexShader = ctx.vs;

		

//manipulate the shaders...

```

	

One can then access some public variables from the vertex and fragment shaders :

```javascript

	

var pinkify = vec4(0.4, 0, 0.4, 1);

	

ctx.fragColor.set(ctx.fragColor.add(pinkify)); //La vie en rose

```

Once every instructions given, one needs to call the *render()* function of the given Context to run the program, so that we have :

```javascript

var tex = new wcv.Texture('#lena'),

	ctx = new wcv.Context('#canvas', tex);

	

var purpleize = vec4(0, 0, 0.7, 1);

ctx.fragColor.set(ctx.fragColor.add(purpleize)); //La vie en rose

	

ctx.render();	

```

	

The above example uses built-in wcv functions to manipulate the underlying shaders, however, if you want and know how to, you can 

edit the programs directly in GLSL ES 2.0 :

```javascript

var tex = new wcv.Texture('#lena'),

	ctx = new wcv.Context('#canvas', tex);

	

	ctx.fs.main.append(

		'vec4 purpleize = vec4(0, 0, 0.7, 1)',

		'fragColor = fragColor + purpleize'

	);

	

	ctx.render();

```

	

both of the above will produce the following result :

[![lena_pinkifify](demo/images/purpleize.png)](https://jsfiddle.net/nathsou/b03391kz/5/)

	

### Going a bit further

##### WebCV comes in with a bunch of useful functions and filters :

- Convolution

A lot of very interesting effects can be applied thanks to a [convolution kernel](https://en.wikipedia.org/wiki/Kernel_(image_processing))

Let's apply the [following kernel](http://homepages.inf.ed.ac.uk/rbf/HIPR2/mean.htm) to our image :

![3 by 3 Mean filter](demo/images/kernel_1.png)

WebCV provides a *convolution* function, ready to use as follows :

```javascript

var kernel = [1/9, 1/9, 1/9, 1/9, 1/9, 1/9, 1/9, 1/9, 1/9];

	

var conv = ctx.convolution(tex, kernel);

	

ctx.fragColor.set(conv);

```

This kernel blurs the image : 

[![blur](demo/images/blur.png)](https://jsfiddle.net/nathsou/b03391kz/1/)

#### Sobel operator

A very common filter in computer vision is the [Sobel Operator](https://en.wikipedia.org/wiki/Sobel_operator) which 

highlights image's contours.

Usage : 

```javascript

var GradientIntensity = ctx.sobel(tex);  //returns a float

ctx.fragColor.set(vec4(vec3(GradientIntensity), 1));

```

Result : 

[![sobel](demo/images/sobel.png)](https://jsfiddle.net/nathsou/b03391kz/4/)

#### Posterize

Usage : 

```javascript

var posterize = ctx.posterize(ctx.fragColor, gamma?, numColors?);  //returns a vec4

ctx.fragColor.set(vec4(vec3(posterize), 1));

```

Result : 

[![posterize](demo/images/posterize.png)](https://jsfiddle.net/nathsou/b03391kz/2/)

#### Combining GLSL & JavaScript

One can easily combine WebCV built-in functions and GLSL code inside the *append()* function call :

```javascript

var G = ctx.sobel(tex);

	

ctx.fs.main.append(

	'vec4 color = vec4(vec3(0), 1)',

	_if(len(distance(ctx.fragColor, vec4(0.75, 0.31, 0.34, 1))).lss(float(0.15)),

		'color = ' + vec4(0, 1, 0, 1)

	)

);

	

ctx.fragColor.set('color');

```

Result :

[![combine](demo/images/combine.png)](https://jsfiddle.net/nathsou/b03391kz/3/)

### Extending WebCV

WebCV lets you add your own functions to the core library