djrobin17 / image-stego-tool Goto Github PK

This might be a python 2 to python 3 thing, but there is a bug in the code that leads to really obviously distorted images (especially obvious where there is a dark background to the image).

Because bin() does not pad with leading zeroes, for small pixel values, you aren't changing only the LSB
e.g. take a pixel value of 17 - the output of bin(17) is 0b10001, and when you take the [2:9] slice of this you get 10001. The new LSB is added to the end of this, so you end up with 100010 or 100011 - which is 34 or 35 so very different from 17!

I'm sure there is a simpler solution, but the below code works :-)

array[p][q] = int(bin(array[p][q])[2:].zfill(8)[:7] + b_message[index], 2)

i.e. convert to binary, remove 0b from start, pad with leading zeroes, then remove LSB

Thanks for the code - was very useful - even if it did drive me slightly nuts debugging my weirdly striped images :-)

I have dont the entire code with opencv and bitarray where i used bitarray instead of using a delimiter to encode and decode the message. I have taken zero padding(null bytes padding) and completed the code with proper exception handling.Can you assign me this issue.