Shivendra Shivani

DCT Based Approach for Tampered Image Detection and Recovery Using Block Wise Fragile Watermarking Scheme

This paper proposes a self- embedding block wise fragile watermarking scheme with tamper detection and content recovery capability. The proposed scheme embeds the shuffled extensive ten bit Recovery data and two bit Authentication data of the image block into the least significant bits (LSB) of its corresponding mapping block.The integrity of a test block is decided by comparing 2×2 non overlapping block of the test block with its corresponding mapping block. Experimental results show that the suggested scheme outperforms conventional self-recovery fragile watermarking algorithms in alteration detection as well as in tamper recovery of the image.

QUANTIZATION-BASED FRAGILE WATERMARKING USING BLOCK-WISE AUTHENTICATION AND PIXEL-WISE RECOVERY SCHEME FOR TAMPERED IMAGE

This paper suggests an efficient fragile watermarking scheme for image content authentication along with altered region restoration capability. In this scheme, image is divided into nonoverlapping blocks of size 2 × 2 and for each block, eight bits for image content recovery data and four bits for authentication data from five most significant bits (MSBs) of each pixel, are generated. These 12 bits are embedded into the least significant bits (LSBs) of the pixels which are placed in its corresponding mapping block. At the receiver end by comparing the recalculated and extracted authentication data, the tampered blocks can easily be identified and using recovery data, one can easily restore the tampered block. Results of experiments demonstrate that the proposed scheme is effective enough for alteration detection as well as tamper recovery of the image.

An effective pixel-wise fragile watermarking scheme based on ARA bits

Potential for exactly tracing maliciously altered pixels is currently desired in image authentication. This paper proposes an efficient pixel-wise fragile watermarking scheme, based on ARA (Authentication Relational & Associative) bits. In this scheme ARA bits are embedded as a watermark into the first three LSBs of each pixel of the host image. On the receiver end, by comparing the extracted LSBs and recalculated ARA bits, one can easily recognize the altered pixels of the host image.

Self-embedding Pixel Wise Fragile Watermarking Scheme for Image Authentication

This paper presents an effective self-embedding pixel wise fragile watermarking scheme for image content authentication along with tampered region localization capability. In this approach the watermark is generated from the five most significant bits (MSBs) of each pixel using three different algorithms and embedded into the three least significant bits (LSBs) of the corresponding pixel in the host image. At the receiver side by comparing the recalculated and extracted LSBs, one can easily identify the tampered pixels of the watermarked image. Results of experiments demonstrate that the proposed scheme has very high fragility and fidelity.

VMVC: Verifiable multi-tone visual cryptography

Traditional k out of n threshold visual cryptography scheme is proposed to hide a secret image into n shares, where only k or more shares can visually reveal the secret image. Most of the previous state of art approaches on visual cryptography are almost restricted in processing of binary images as secret, which are inadequate for many applications like securely transmission of medical images(Store and Forward Telemedicine), forensic images etc. In this paper, a new Verifiable Multi-toned Visual Cryptography (VMVC) scheme is proposed to securely transmit the confidential images on web. Proposed approach also provides cheating prevention, since each pixel of shares contains a self embedding verifiable bit for integrity test of that pixel. Many existing approaches are suffering from many unnecessary encryption constraints like random shares, codebook requirement, contrast loss etc, which all are successfully addressed in proposed approach. Some comparisons with previously proposed methods are also made. Experimental results and analysis are used to prove the efficiency of proposed approach.

Progressive Visual Cryptography with Unexpanded Meaningful Shares

The traditional k-out-of-n Visual Cryptography (VC) scheme is the conception of “all or nothing” for n participants to share a secret image. The original secret image can be visually revealed only when a subset of k or more shares are superimposed together, but if the number of stacked shares are less than k, nothing will be revealed. On the other hand, a Progressive Visual Cryptography (PVC) scheme differs from the traditional VC with respect to decoding. In PVC, clarity and contrast of the decoded secret image will be increased progressively with the number of stacked shares. Much of the existing state-of-the-art research on PVC has problems with pixel expansion and random pattern of the shares. In this article, a novel scheme of progressive visual cryptography with four or more number of unexpanded as well as meaningful shares has been proposed. For this, a novel and efficient Candidate Block Replacement preprocessing approach and a basis matrix creation algorithm have also been introduced. The proposed method also eliminates many unnecessary encryption constraints like a predefined codebook for encoding and decoding the secret image, restriction on the number of participants, and so on. From the experiments, it is observed that the reconstruction probability of black pixels in the decoded image corresponding to the black pixel in the secret image is always 1, whereas that of white pixels is 0.5 irrespective of the meaningful contents visible in the shares, thus ensuring the value of contrast to alwasys be 50%. Therefore, a reconstructed image can be easily identified by a human visual system without any computation.

A chaotic map based DCT-SVD watermarking scheme for rightful ownership verification

A chaotic map-based hybrid watermarking scheme incorporating the concepts of the Discrete Cosine Transform (DCT) and exploiting the stability of the singular values has been proposed here. Homogeneity Analysis of the cover image has been done to chalk out appropriates sites for embedding and thereafter, a reference image has been obtained from it. The singular values of the reference image has been modified for embedding the secret information. The Chaotic map based scrambling enhances the security of the algorithm as only the rightful owner possessing the secret key, could retrieve the actual image. Comprehensive set of attacks has been applied and robustness tested with the Normalized Cross Correlation (NCC) and Peak Signal to Noise Ratio (PSNR) metric values. High values of these metrics signify the appropriateness of the proposed methodology.

Multi secret sharing with unexpanded meaningful shares

Traditional Visual Cryptography (VC) facilitates a technique to protect only one secret image using single set of shares. Recent researches enhance the capabilities of traditional VC by providing the feature of Multi Secret Sharing (MSS), where more than one secret image can be protected at a time. In MSS different secret images are revealed by the stacking of same set of shares at different angles. Most of the existing state of art researches on MSS have common problem of pixel expansion and random pattern of the shares. Due to pixel expansion, there is wastage of the storage space and transmission time, moreover random pattern of the shares increases the vulnerability for cryptanalysis. In this paper a novel Multi Secret Sharing scheme with unexpanded as well as meaningful shares has been proposed to protect two secret images at a time. In the proposed approach the recovery probability of black pixels of the secret images in the decoded images is always 1 while that of white pixels, it is 0.25. Therefore the contrast of the decoded images is obtained as 25 % which is same as in most of the earlier researches with pixel expansion & random shares. Experiments confirm that all meaningful shares fulfill the contrast and security conditions. Secret images can be easily decoded by only human visual system without any computation at receiver end.

Self authenticating medical X-ray images for telemedicine applications

Telemedicine focuses on sharing medical images over network among doctors for better consultation. Hence medical images must be protected from unwanted modifications (intentional/unintentional) over network. Digital watermarking can be used for this purpose. ROI based watermarking hide’s ROI (Region of Interest) information in NROI (Non-Interested Regions of Interest) regions in general. Most of the existing watermarking methods used in telemedicine are block based. Main drawback of block based watermarking is exact localization of exact tampered pixels, as even if single pixel of a block is altered the whole block is detected as tampered. Thus this paper proposes a pixel based watermark technique so that alteration can be detected exactly. This method works in three phases: first, separates the ROI and NROI, second some information depending on the relative size of ROI and NROI the ROI is embedded in NROI using LSBs of ROI. Third, a tamper detection algorithm is proposed for detecting the tampered pixels exactly. Performance of proposed method when compared with state-of-art block based approaches is found to be best having 33db of PSNR value and 80% of accuracy on an average.

XOR based continuous-tone multi secret sharing for store-and-forward telemedicine

Traditional k out of n visual cryptography scheme has been proposed to encrypt single secret image into n shares where only k or more shares can decode the secret image. Many existing schemes on visual cryptography are restricted to consider only binary images as secret which are not appropriate for many important applications. Store-and-Forward telemedicine is one such application where medical images are transmitted from one site to another via electronic medium to analyze the patient’s clinical health status. The main objective of Store-and-Forward telemedicine is to provide remote clinical services via two-way communication between the patient and the healthcare provider using electronic medical image, audio and video means. In this paper, a new XOR based Continuous-tone Multi Secret Sharing scheme suitable for store-and-forward telemedicine is proposed to securely transmit the medical images. It also eliminates basic security constraints of VC like pixel expansion in shares/recovered secret images, random pattern of shares, explicit codebook requirement, lossy recovery of secret and limitation on number of secret and shares. Proposed approach is n out of n multi secret sharing scheme which is able to transmit n secret images simultaneously. All secrets could be revealed only after some computations with all n shares and one master share. Master share has been created with the secret key at encoding phase and it can be regenerated at the time of decoding using same secret key. Here all shares are meaningful in continuous-tone which may provide confidentiality to medical images during transmission. Proposed approach not only preserves all basic characteristics of traditional VC but also increases the capacity of secret image sharing. From the experiments we found that irrespective of visible contents of the shares, the probability of getting back the pixel values of respective original secret images at the receiver end is very high.