What's special about this image?

Question

There's something very special hidden in this image. You will definitely know when you find it.

(I know Stack Exchange lossily compresses images but I've checked and no pixels have been altered.)

Finding traces of the secret might not be that hard. The ideal answer will point to the Stack Exchange question that I took data from to help generate this image (and possibly point to another that can help "decode" it).

Hint:

I can reveal the secret with this 20 line Python program. I've replaced the most telling bits with ???:

from PIL import Image
i = Image.open('hidden.png')
d = i.load()
x, y = ???, ???
while ???:
    oldX, oldY = x, y
    if ???:
        ???
    elif ???:
        ???
    elif ???:
        ???
    else:
        ???
    d[oldX, oldY] = (0, 0, 0)
for x in range(i.size[0]):
    for y in range(i.size[1]):
        if ???:
            d[x, y] = (255, 255, 255)
i.save('revealed.png')

@Bobson Not really. You may find a png that is similar to the original but you won't find an exact original because I had to remake it to make sure it was pixel perfect. — Calvin's Hobbies, Mar 16 '15 at 19:20

score 16 · Answer 1 · answered Mar 16 '15 at 04:30

16

Putting it through a magic eye finder gives a faint image of a smiling face, perhaps the Mona Lisa.

answered Mar 16 '15 at 04:30

xnor

26,756
4
85
144

4

Wow. That algorithm has no (intentional) relation to the image. I'm surprised it shows so much. – Calvin's Hobbies Mar 16 '15 at 04:46
17

Wait? That is not the solution? You are saying this is meant to represent something else, but by chance also contains Mona Lisa? – Lars Ebert Mar 16 '15 at 08:28
Well it depends on how that website works. It might output some random things. – Vladimir Cravero Mar 16 '15 at 13:23
This code golf question could be what Calvin is referencing, but I don't understand what he wants to do with it. (It is #5 on the highest vote list.) – Len Mar 16 '15 at 18:30
@LarsEbert Mona Lisa is the hidden thing, but there's a simpler way to reveal her in a much crisper form. – Calvin's Hobbies Mar 16 '15 at 19:17
@Len The question I took data from to generate the image is not my own (the "decoder" however...) – Calvin's Hobbies Mar 16 '15 at 19:18
Well, this result is achieved only playing with saturation and colors. I've resized the image since it's more visible when smaller. – Narmer Mar 17 '15 at 09:16

score 8 · Accepted Answer · answered Mar 17 '15 at 08:14

This image represents instructions for a snake that fills the visited pixels with black paint.

Red color means "go right"

Aqua color means "go left"
Blue color means "go up"
Green color means "go down"

The snake starts at the center of the image. (Edit: looks like it can start anywhere but the borders. The images below were produced using center as the starting point.)
The snake stops as soon as it proceeds out of the image boundaries.
After we're done, we fill remaining unvisited pixels with white paint.

What is interesting is what happens if the snake visits a block that was already visited: it turns out that no matter if we go left, right, down or up, we still see the result pretty clearly.

Going left:
Going down: Going right: Going up:

#!/bin/python
from PIL import Image
colors = [(255, 0, 0), (0, 255, 255), (127, 0, 255), (127, 255, 0)]
i = Image.open('hidden.png')
d = i.load()
w, h = i.size
x, y = w >> 1, h >> 1
while x >= 0 and y >= 0 and x < w and y < h:
    oldX, oldY = x, y
    if d[x,y] == colors[0]:
        x += 1
    elif d[x,y] == colors1:
        x -= 1
    elif d[x,y] == colors2:
        y += 1
    elif d[x,y] == colors3:
        y -= 1
    elif d[x,y] == (0,0,0):
        y -= 1 #or whatever
    else:
        raise Exception('Bad color: ' + str(d[x,y]))
    d[oldX, oldY] = (0, 0, 0)
for x in range(w):
    for y in range(h):
        if d[x, y] != (0, 0, 0):
            d[x, y] = (255, 255, 255)
i.save('revealed.png')

This is essentially it :D The pixels that aren't part of the snake are actually random colors (of the 4) so there's no guarantee that they will lead to the snake. Good job! (Though I'll still wait for someone to figure out which two SE questions I hinted at.) — Calvin's Hobbies, Mar 17 '15 at 08:27
You shouldn't create an answer which is ALL spoiler tags without any significant text outside. — Marco Bonelli, Mar 18 '15 at 12:30

What's special about this image?

2 Answers2