16-01-2013, 09:06 AM
Data Hiding and Authentication Watermarking
Abstract:
In image authentication watermarking, hidden data is inserted into an image to detect any accidental or malicious image alteration. In the literature, quite a small number of cryptography based secure authentication methods are available for binary images. In a cryptography based authentication watermarking, a message authentication code (or digital signature) of the whole image is computed and the resulting code is inserted into the image itself. This paper proposes a new authentication watermarking method for binary images. The main idea is to use the prioritized sub-blocks by pattern matching scheme to embed the code. Shuffling is applied before embedding to equalize the un even embedding capacity. It detects any alteration while maintaining good visual quality for all types of binary images. The security of the algorithm lies only on the secrecy of a secret or private keys used.
Introduction
Data hiding represents a class of processes used to embed data, such as copyright information into various forms of media such as image, audio, or text with a minimum amount of perceivable degradation to the “host” signal; its goal is not to restrict or regulate access to the host signal, but rather to ensure that embedded data remains inviolate and recoverable.
A watermarking technique makes use of a data hiding scheme to insert some information in the host image, in order to make an assertion about the image later. In this paper, data hiding scheme simply means the technique to embed a sequence of bits in a still image and to extract it after wards.
Water marking techniques can be classified as either “robust” or “fragile.” Robust water marks are useful for copyright and ownership assertion purposes. They cannot be easily removed and should resist common image manipulation procedures. On the other hand, fragile watermarks (or authentication watermarks) are easily corrupted by any image processing procedure. However, watermarks for checking the image integrity and authenticity can be fragile because if the watermark is removed, the watermark detection algorithm will correctly report the corruption of the image .
Data Hiding and Authentication Watermarking
In the literature, there are many authentication watermarking techniques for continuous-tone images .Also, there are many techniques for data hiding in binary and halftone images. There are many papers on data hiding in binary images but there is no authentication water marking.
How ever, quite a small number of secure authentication watermarking techniques are available for binary and halftone images. Only some techniques used the cryptography-based secure authentication water marking. It was given for dispersed-dot halftone images but the visual quality for a binary image is poor. The ratio of black versus white pixels is used. Although the algorithm aims at robustly hiding information in binary image, it is not secure enough to be directly applied for authentication or other fragile use.
Some times PWLC does not correctly extract the hidden data, and fails to recover perfectly the original cover image. In summary, these previously proposed approaches either cannot be easily extended to other binary images, or can only embed a small amount of data.
In secure authentication watermarking using some data hiding technique for binary image, one must compute a hashing function of the binary image F, obtaining the hash value H = H(F). After encryption, it becomes MAC /DS. This MAC/DS must be inserted into F itself, obtaining the marked image F′.
The SAWT-PM
In SAWT-PM, only a few pixels are modified and the positions of sub-blocks containing those pixels are known both in the insertion and extraction phases. SAWT-PM flips only low-visibility pixels to hide the information and consequently the watermarked image has excellent visual quality and do not have salt-and-pepper noise.
Finding Ready Blocks
The original image is represented as F. It is partitioned into m × n sub-blocks Fi. In each sub-block only the Middle pixel is used to hide the information. But, not all the sub-blocks are used for hiding the information. Among the 512, 3 × 3 sub-blocks several blocks were rejected due to various reasons like, high visibility, Non reversible at receiver side etc. Only 120 patterns are considered to be valid for data hiding.
The Figure 1 depicts the valid patterns. The hatched middle pixel may either be black or white. The change of middle pixel is less noticeable. Mirrors, transposes and reverses of the patterns are also used for hiding the information.
Each 3 × 3 sub-block is checked against various constraints to identify that whether it matches with any of the valid 120 patterns or not. If a sub-block matches with any of the valid patterns, it is a Ready Block.
Authentication Signature
Let k be the length of the adopted AS. To insert k bits of AS, it needs k, m × n Ready Blocks in the image. The image is divided into two regions: The first region is small in size, called AS Region (ASR) where AS will be inserted and the second region is the original image excluding ASR, called Non-AS Region (NASR) from where AS will be computed. Using this technique, the image authenticity can be verified at the receiver side.
Once the Ready Blocks are found, it is clearly understood that the middle pixels of all the Ready Blocks forms the ASR. So, before calculating the hash value of the image, the ASR is made to be darkened. The new Image is referred as (ZASR –Zeros inserted at AS Region).The ZASR is actually the whole image with zeros in ASR. Now, the hash value is calculated for ZASR, encrypted using the secret- or public key ks, obtaining the MAC/DS and is inserted in ASR .
At the receiving side, the receiver uses the same technique to identify the Ready Blocks. ASR is separated, encrypted AS inserted in that are retrieved and decrypted using the secret- or public key ks, obtaining the AS. Then in the received original image, middle pixels of ASR are made to be darkened. It is referred as ZASR*. Compute H(ZASR*) and if H (ZASR*) =AS, then image integrity is verified. Otherwise, image has been modified or a wrong key was used.
Data Hiding Scheme
This technique ensures that for any pixel that is modified in the host image, its visibility is very less. Thus, the existence of secret information in the host image is difficult to detect.
The data hiding scheme is very simple. In each block, the position of the pixel to be used for hiding the information is fixed, that is, the middle pixel. Hence the complexity is not in identifying the position, but in identifying the sub-blocks that are ready to hide the information (Ready Blocks).
Since the middle pixel is used to hide the information always, another level of security is introduced by shuffling of the Ready Blocks. This provides another advantage also. It distributes the hidden information in all parts of the image. It is an efficient and effective tool to equalize uneven embedding capacity.
The Ready Blocks are shuffled and permuted order is generated randomly. A secret key, called Shuffling Key shared by the sender and the receiver, is used as the seed for pseudo random number generator.