Twitter @_cjwallace
outputPixels = {
let inputPixels = Array.from(input.data)
for (let mask of segmentation) {
inputPixels = maskImage(inputPixels, mask)
}
let output = new ImageData(WIDTH, HEIGHT)
output.data.set(inputPixels)
return output
}
input = {
let ctx = DOM.canvas(image.width, image.height).getContext("2d");
ctx.drawImage(image, 0, 0);
return ctx.getImageData(0, 0, image.width, image.height)
}
function maskImage(pixels, mask) {
"Mask the image by replacing each pixel in the person segmentation with a randomly selected nearby pixel"
let maskPixels = mask.data
// iterate over pixels
for (let ix = 0; ix < maskPixels.length; ix++) {
let [x_, y_] = ixToCoords(4*ix, mask.width)
// mask pixel in image if part of target segment
if ((target == "PEOPLE" && (maskPixels[ix] != -1)) ||
(target == "HEADS" && maskPixels[ix] in [0,1])) {
pixels = maskPixel(pixels, x_, y_)
}
}
return pixels
}
function maskPixel(pixels, x, y) {
"Replace a pixel with a random nearby pixel."
let [shiftedX, shiftedY] = randomNoise(x, y)
// pixels have four indices corresponding to rgba channels
for (let j=0; j<4; j++) {
pixels[coordsToFirstIx(x, y, WIDTH)+j] = pixels[coordsToFirstIx(shiftedX, shiftedY, WIDTH)+j]
}
return pixels
}
function coordsToFirstIx(x, y, width) {
return 4 * (x + y * width)
}
function ixToCoords(ix, width) {
let x = Math.floor(ix % (4 * width) / 4)
let y = Math.floor(ix / (4 * width))
return [x, y]
}
function randomNoise(x, y) {
let scale = NOISE
// initialization and while to avoid boundary effects of shifting to outside the image
let shiftX = -1
let shiftY = -1
while (shiftX < 0 || shiftX >= WIDTH || shiftY < 0 || shiftY >= HEIGHT ) {
shiftX = x + randInt(-scale, scale)
shiftY = y + randInt(-scale, scale)
}
return [shiftX, shiftY]
}
function randInt(min, max) {
return Math.floor(Math.random() * (max - min + 1) ) + min
}
One platform to build and deploy the best data apps
Experiment and prototype by building visualizations in live JavaScript notebooks. Collaborate with your team and decide which concepts to build out.
Use Observable Framework to build data apps locally. Use data loaders to build in any language or library, including Python, SQL, and R.
Seamlessly deploy to Observable. Test before you ship, use automatic deploy-on-commit, and ensure your projects are always up-to-date.