Image Authentication by Content Preserving Robust Image Hashing Using Local and Global Features

Abstract

Image hashing technique constructs a short sequence from the image to represent its contents. This method proposes an image hash which is generated from Haralick and MOD-LBP features along with luminance and chrominance, which are computed from Zernike moments. Sender generates the hash from image features and attaches it with the image to be sent. The hash is analyzed at the receiver to examine the authenticity of the image. The method detects image forgery and locates the forged regions of the image. The proposed hash is robust to common content preserving modifications and sensitive to malicious manipulations.

Fast Legendre Moments in Securing Digital Image

“…Digital images are used in a wide range of applications for the past two decades and as a result of that many image editing applications challenges the security and integrity of images. There are many ways a hash can be generated by extracting local and global features [4],fixed point theory [7], watermarking methods [3] [8], NMF [9] [12], using fourier mellin transformation [12],DFT coefficients, radon transform co efficient as hash code [15],image histogram [13],DWT coefficients [3], [21] and much more. In recent work combining two or more hashing techniques to generate the hash code is also most common to make advantage of different hashing techniques [20].…”

“…Moments are set of values that best describes the content of the image [1] . Many existing system uses Zernike moments for extracting the global features of the image [4] the reason is that Zernike moment is computationally inexpensive than legendre moments.…”

Adaptive CSLBP compressed image hashing

“…Following represents various global and local features pairs for content change location locally as well as globally. DWT-SVD and Saliency object detection using spectral residual model; Projected Gradient Non-negative Matrix Factorization (PGNMF), ring partition and saliency detection; Zernike moment and Salient point detection; Zernike moment and Haralick local features; Zernike moments, MOD-LBP and Haralick texture features; Invariant moments from Radon coefficients and statistical measures from Radon coefficients; DCT coefficients of Watson's visual model and SIFT key points; Color vector angle and Salient edge points [5][6][7][8][9][10][11][12].…”

Combining Image Security and Authentication Using Sift and Zernike Moments

“…The content of the image is represented as a small code called Hash code [1]. Hash code can be generated by extracting local and global features [2], Watermarking methods [3] [4], Transform co-efficient [5] [6], and much more. In recent work combining two or more hashed techniques to generate the hash code is also most common to take advantage of different hashing techniques [7].…”

“…A proper subset of moments is always the best choice that describes the exact content of image. Zernike moments have an orthogonal basis functions and are used as Global image features in most authentication system [2].Lv et al [12] explained about local feature points and shape context by using SIFT as feature extraction method. SIFT is used to select the most stable key points from the image and it is embedded to generate image hash.…”