11

11 - Oversized Gifts

Description

To ensure that Santa does not have to carry such a heavy load, our elves are always trying to shrink the gifts as much as possible. New technologies are constantly being developed in our laboratories. Unfortunately, an incident occurred during a test, when restoring the original size, an error occurred and now we are no longer able to achieve the original size.

Solution

For this challenge we are given a huge PNG file. Opening it isn't simply possible since it wouldn't even fit into RAM. We have to scale it down somehow. I found two different solutions to this challenge.

First Solution using VIPs

One solution uses the image processing library vips. We can simply call the following command to resize the image and then read the QR code:

vips resize 72d85b7f-4325-432e-93ff-cfdc019306c6.png out.png 0.0078125 --vips-progress

Second Solution using Java

Alternatively, we can just take the pixel data using a library. In my case, I took PNGJ a java library for PNG decoding. From a hint that was published during the competition, we know that the original QR code (before scaling) was generated using qrencode. This means that the original image was of size 111x111. The scaled version was of size 909312x909312 and the scaling factor was 8192.

With this in mind, we can extract every 8192-th pixel in a grid-like fashion:

PngReader pngReader = new PngReaderByte(new File("72d85b7f-4325-432e-93ff-cfdc019306c6.png"));
BufferedImage image = new BufferedImage(111, 111, BufferedImage.TYPE_INT_RGB);
int outputRow = 0;

for (int row = 0; row < pngReader.imgInfo.rows; row += 8192) {
    IImageLine line = pngReader.readRow(row);
    int outputCol = 0;

    for (int i = 0; i < ((ImageLineByte) line).getScanline().length; i += (4 * 8192)) {
        byte r = ((ImageLineByte) line).getScanline()[i];
        byte g = ((ImageLineByte) line).getScanline()[i + 1];
        byte b = ((ImageLineByte) line).getScanline()[i + 2];
        byte a = ((ImageLineByte) line).getScanline()[i + 3];
        int rgb = (r << 16 | g << 8 | b);

        image.setRGB(outputRow, outputCol, rgb);
        outputCol++;
    }

    outputRow++;
    ImageIO.write(image, "PNG", new File("done.png"));
}

pngReader.end();