bmv

Block Mean Value Based Hashing, Method 1, is a robust hash function which:

1. Grey scale and normalises image size;
2. Divides the image into non-overlapped blocks i;
3. Encrypts the indices of block sequences i (skipped in this implementation);
4. Calculates the mean (Mi) of each block, based on pixel value, and obtain Median (Md) of calculated means.
5. Normalize mean values into binary hash form: h(i) = 0 if Mi < Md or 1 if Mi >= Md.

The implementation is based on Method 1 (note that the encryption step has been skipped) orginally described by:

B. Yang, F. Gu and X. Niu, "Block Mean Value Based Image Perceptual Hashing," 2006 International Conference on Intelligent Information Hiding and Multimedia, 2006, pp. 167-172, doi: 10.1109/IIH-MSP.2006.265125.

Usage

Run the following commands in your Python (3.x) environment:

git clone https://github.com/blockpass-identity-lab/bmv
pip install -r requirements.txt
python bmv.py

Example output:

=============================================================================================================================
malicious.png binary hash: [1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0];
=============================================================================================================================
modified1.png binary hash: [1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0]; hd: 2;  match: True
modified2.png binary hash: [1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0]; hd: 0;  match: True
modified3.png binary hash: [1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0]; hd: 0;  match: True
benign.png binary hash:    [0, 0, 0, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1]; hd: 18; match: False
benign2.jpg binary hash:   [0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 0]; hd: 16; match: False

Table of images and results:

Name	Image (resized to 128x128)	Hamming Distance	Match
malicious.png
modified1.png		2	True
modified2.png		0	True
modified3.png		0	True
benign.png		18	False
benign2.jpg		16	False