JavaScripting

The definitive source of the best
JavaScript libraries, frameworks, and plugins.


  • ×

    Image Match - Project to allow compare different images pixel by pixel with good performance
    Filed under  › 

    • 🔾3%Overall
    • 81
    • 196.2 days
    • 🕩7
    • 👥1

    IM.js

    IM.js is a class to compare images in canvas at pixel level.

    Changelog

    1.3

    • Refactored to reduce final size
      • Reduction:
        • Minimized:
          • 25% reduction - 2.880 bytes --> 2.157 bytes
        • Gzipped:
          • 14% reduction - 1.442 bytes --> 1.215 bytes

    1.2.1

    • Changed setDiff to showDiffInCanvas
    • Changed setPercentageDiff to setTolerance

    1.2.0

    • Added setPercentageDiff method to set the tolerance percentage
    • Changed algorithm to see the differences in canvas.
      • Now all the matching pixels will be showed transparent and the not matching pixels will be showed as they are showed in the image that is different. The canvas is bordered to show its bounds.

    1.1.0

    • Added setDiff method to get diff image in canvas ** All the matching pixels will be showed in black and the not matching pixels will be showed inverting the color.

    1.0.0

    • Using compare you can create image in canvas at the same time to make the compare or compare canvas with images without the need of create them.
    • The result of compare is sent to the callbacks that tell if the image are equals or not.
    • The first canvas with image info different from the previous images is marked with one red border.

    Previous to IM

    Tools:

    • Desktop tools
    • Javascript classes that compare MD5 or base64 result of enconding the image.

    These Javascript classes only compare that the image is not the same. The problem is that if you have some image with the same pixel info the result could be different and return false.

    Imagine that you have an image with different compression rate and you want to know if the result is the same.

    Then my only solution was to check images at pixel info level.

    Description

    IM checks all your images using canvas at pixel level.

    When I say compare images at pixel level is compare the image pixel to pixel.

    Yes it sounds hard! and it is! ;)

    I expent a lot of time trying to do it, but some implementations collapse my browser. I try the same but using workers with the same result.

    After some unsuccessful attempts I found the final solution that is that now is implemented in IM.

    In the examples you can see how it can test 9 images (300px x 250px) in less than 500 ms.

    • Different images example only tooks 296ms.

    • Equals images example only tooks 449ms

    Browser support:

    • IE9 and best
    • Mozilla Firefox
    • Google Chrome
    • Safari
    • Konqueror
    • Opera

    Benefits and usages:

    The idea of create IM was to automatize the image comparison.

    I use it in my testing environment.

    I use Selenium to open the same url in differents browsers and get a screenshot of each one.

    I save it with the same name on one folder, then I open one tool where I implemented IM and use it to save the result in the C.I.(Continuous integration) server. All is done in no more than 10 seconds.

    API documentation

    Examples to see for yourself!

    Usage

    Before using it:

    Insert in your code:

    <script type="text/javascript" src="/path/to/your/js/libs/IM.js"></script>
    

    Set debug mode

    IM.setDebug(boolean);
    

    If you set debug mode you will be able to see how many time tooks the comparing, and you will see a border in the canvas that triggers the fail if it fails.

    Set asynchronous mode

    IM.setAsynchronous(boolean);
    

    IM use synchronous loops , by default, when checking the image data in canvas.

    If for some reason in some environment you see any problem of performance that blocks your browser, you can set the asynchronous mode to true, it's a bit slower but not blocking method.

    Create a image object config

    When using the create canvas and image mode you will need to create config image objects.

    new IM.image('path/to/images/image.ext', width, height);
    

    Compare creating canvas with images

    When you use this mode you will create a canvas for image

    IM.compare(DOM_Element,
    [
        new IM.image('path/to/images/image.ext', 100, 100),
        new IM.image('path/to/images/image2.ext', 100, 100),
        new IM.image('path/to/images/image.ext', 100, 100),
        new IM.image('path/to/images/image2.ext', 100, 100)
    ],
    function success(aCanvas, nElapsedTime) {
        // Code on success. All images have the same pixel info.
    },
    function fail(oCanvas, nElapsedTime) {
        // Code on failing. Any image is different from others
    });
    

    Compare using existent canvas

    When you use this mode you will need to have the canvas with images before comparing. This allows you to compare images in an existent environment.

    IM.compare([
        document.getElementById("canvas_1"),
        document.getElementById("canvas_2"),
        document.getElementById("canvas_3"),
        document.getElementById("canvas_4")
    ],
    function success(aCanvas, nElapsedTime) {
        // Code on success. All images have the same pixel info.
    },
    function fail(oCanvas, nElapsedTime) {
        // Code on failing. Any image is different from others
    });
    

    Create canvas with diff result

    This could be used to see the result of differences in images in one canvas. Only needs one DOM element where insert the canvas. You can stop diff if you set it to false;

    IM.showDiffInCanvas(document.getElementById("canvasDiff"));
    

    Set tolerance percentage

    Default value of tolerance is 100%. You can modify the percentage to mark an image ok or ko.

    IM.setTolerance(80);
    

    Documentation

    API documentation

    Examples to see for yourself!

    License

    IM.js is licensed under the MIT license.

    Show All