skeletonization.js library

GPU-powered image skeletonization for the web.

Based on orginal C++ code by Zhang-Suen.

With GPU support from gpu.js.

Uses OpenCV.js.

Checkout the live demo at skeletonization-js.glitch.me.

Usage

In your HTML file:

<!-- Include OpenCV.js -->
<script src="https://docs.opencv.org/3.4/opencv.js"></script>

<!-- Include gpu.js -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/gpu.js/1.10.4/gpu.min.js"></script>

<!-- Include skeletonization.js -->
<script src="https://skeletonization-js.glitch.me/skeletonization.js"></script>

In your Javascript file:

// Usage with HTML Canvas:

// setup skeletonization:
skeletonization.setup(
  256, // output width
  256  // output height
);

// skeletonize a canvas element specified by its id:
skeletonization.skeletonize("inputCanvasId", {
  preprocess: true, // whether or not to preprocess the
                    // source image (blur, threshold, etc.)
    blur: 5,        // if preprocess, radius of blurring to apply
    threshold: 128, // if preprocess, binary threshold to apply
    invert: false,  // if preprocess, invert the image (foreground
                    // should be white, background should be black)
  outputCanvasId: "outputCanvasId", // id of canvas on which output
                                    // will be displayed
  bbox: [0,0,256,256], // bounding box (xmin,ymin,xmax,ymax) of
                       // the region to apply skeletonization,
                       // leave undefined to apply to whole image
})

You can also skeletonize cv::Mat, e.g. in an OpenCV.js project:

// Usage with OpenCV.js:

// setup skeletonization
skeletonization.setup(
  256, // output width
  256  // output height
);

// read the image with OpenCV.js
var im = cv.imread("inputCanvasId");

// ... (custom preprocessing)

// skeletonize the cv::Mat
// skeletonize modifies the input cv::Mat,
// so save a copy if you still need the original
skeletonization.skeletonize(im, {
  preprocess: true, // whether or not to preprocess the
                    // source image (blur, threshold, etc.)
    blur: 5,        // if preprocess, radius of blurring to apply
    threshold: 128, // if preprocess, binary threshold to apply
    invert: false,  // if preprocess, invert the image (foreground
                    // should be white, background should be black)
  bbox: [0,0,256,256], // bounding box (xmin,ymin,xmax,ymax) of
                       // the region to apply skeletonization,
                       // leave undefined to apply to whole image
})

// ... (custom postprocessing)

// display the result
cv.imshow("outputCanvasId",im);