31-07-2014, 10:11 AM
Image Hashing
Image Hashing.pptx (Size: 233.36 KB / Downloads: 12)
Introduction
Due to the ever increasing digitalization, the authentication of multimedia content is becoming more and more important.
Authentication:- Verifying the originality of an object.
The authentication depends heavily on the type of the object.
In cryptography, hash functions are used as digital signatures to authenticate the message being sent.
When authenticating an executable file or text, it is important that every single bit exactly matches the original data.
Multimedia data(eg: image) can allow for lossy representations with graceful degradation.
Hash function
It is a computationally efficient one-way function mapping an object of arbitrary length, to binary strings of some fixed length called hash values.
Let I be a variable size message and h be its fixed size hash value. Then the hash function H can be defined as,
h = H ( I )
Hash functions can be divided into two categories.
Cryptographic Hash functions.
Perceptual hash functions.
Distance/Similarity Functions for Perceptual Hashes
A perceptual hash function calculates closer hash values for similar media objects and distinct hash values for dissimilar objects.
To compare two perceptual hashes appropriate measures must be used.
The most often used distance metrics are the Bit Error Rate (BER), Euclidean distance and Peak of Cross Correlation (PCC).
The first two measure the distance between two hash values, whereas the latter measures the similarity between two hash values.
Conclusion
It is clear from these results that it is easy to put a threshold, so that we can differentiate, images which undergone manipulations like compression, filtering etc with other images.
The 2 image hashing methods presented here are invariant almost all distortions except rotation. Use of rotation invariant transformations like Fourier-Mellin Transform or Radon transform, these methods can be modified to achieve good performance for rotation also.