提取EOF后隐藏的图像

Question

For a little lesson in steganography, I am appending an image to another image file like so:

my_image = open(output_image_path, "wb")
my_image.write(open(visible_image, "rb").read())
my_image.write(open(hidden_image, "rb").read())
my_image.close()

Now I want to extract the hidden image again. How would I do this? I tried with PIL by reading the image or by reading in the file as a bytes stream and then converting it, but I only get the visible image.

In case it matters, I should specify that all images are saved in.jpg format

Answer 1

Ok got it, this is how to show the hidden image:

from io import BytesIO
import cv2
from PIL import Image

with open(my_image, 'rb') as img_bin:
    buff = BytesIO()
    buff.write(img_bin.read())
    buff.seek(0)
    bytesarray = buff.read()
    img = bytesarray.split(b"\xff\xd9")[1] + b"\xff\xd9"
    img_out = BytesIO()
    img_out.write(img)
    img = Image.open(img_out)
    img.show()

Answer 2

I was preparing an answer, and just while typing you added your solution. Nevertheless, here's my version, capable extracting all images stored in the output image:

from io import BytesIO
from PIL import Image

# Create "image to the world"
my_image = open('to_the_world.jpg', 'wb')
my_image.write(open('images/0.jpg', 'rb').read())   # size=640x427
my_image.write(open('images/1.jpg', 'rb').read())   # size=1920x1080
my_image.write(open('images/2.jpg', 'rb').read())   # size=1920x1200
my_image.close()

# Try to read "image to the world" via Pillow
image = Image.open('to_the_world.jpg')
print('Read image via Pillow:\n{}\n'.format(image))

# Read "image to the world" via binary data
image = open('to_the_world.jpg', 'rb').read()

# Look for JPG "Start Of Image" segments, and split byte blocks
images = image.split(b'\xff\xd8')[1:]

# Convert byte blocks to Pillow Image objects
images = [Image.open(BytesIO(b'\xff\xd8' + image)) for image in images]
for i, image in enumerate(images):
    print('Extracted image #{}:\n{}\n'.format(i+1, image))

Of course, I also used the binary data of the output image, and split the binary data using the JPEG file format structure , the "Start of Image" segment FF D8 to be precise.

For the set of images, I used, the output would be the following:

Read image via Pillow:
<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=640x427 at 0x1ECC333FF40>

Extracted image #1:
<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=640x427 at 0x1ECC333FF10>

Extracted image #2:
<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=1920x1080 at 0x1ECC37D4C70>

Extracted image #3:
<PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=1920x1200 at 0x1ECC37D4D30>

----------------------------------------
System information
----------------------------------------
Platform:      Windows-10-10.0.16299-SP0
Python:        3.9.1
PyCharm:       2021.1.1
Pillow:        8.2.0
----------------------------------------