S. Srinivas Kumar

A Robust Image Watermarking Scheme using Singular Value Decomposition

This paper presents a robust image watermarking scheme for multimedia copyright protection. In this work, host image is partitioned into four sub images. Watermark image such as ‘logo’ is embedded in the two of these sub images, in both D (singular and diagonal matrix) and U (left singular and orthogonal matrix) components of Singular Value Decomposition (SVD) of two sub images. Watermark image is embedded in the D component using Dither quantization. A copy of the watermark is embedded in the columns of U matrix using comparison of the coefficients of U matrix with respect to the watermark image. If extraction of watermark from D matrix is not complete, there is a fair amount of probability that it can be extracted from U matrix. The proposed algorithm is more secure and robust to various attacks, viz., JPEG2000 compression, JPEG compression, rotation, scaling, cropping, row-column blanking, row-column copying, salt and pepper noise, filtering and gamma correction. Superior experimental results are observed with the proposed algorithm over a recent scheme proposed by Chung et al. in terms of Bit Error Rate (BER), Normalized Cross correlation (NC) and Peak Signal to Noise Ratio (PSNR).

CONTENT BASED IMAGE RETRIEVAL USING EXACT LEGENDRE MOMENTS AND SUPPORT VECTOR MACHINE

Content Based Image Retrieval (CBIR) systems based on shape using invariant image moments, viz., Moment Invariants (MI) and Zernike Moments (ZM) are available in the literature. MI and ZM are good at representing the shape features of an image. However, non-orthogonality of MI and poor reconstruction of ZM restrict their application in CBIR. Therefore, an efficient and orthogonal moment based CBIR system is needed. Legendre Moments (LM) are orthogonal, computationally faster, and can represent image shape features compactly. CBIR system using Exact Legendre Moments (ELM) for gray scale images is proposed in this work. Superiority of the proposed CBIR system is observed over other moment based methods, viz., MI and ZM in terms of retrieval efficiency and retrieval time. Further, the classification efficiency is improved by employing Support Vector Machine (SVM) classifier. Improved retrieval results are obtained over existing CBIR algorithm based on Stacked Euler Vector (SERVE) combined with Modified Moment Invariants (MMI).

Medical Image Segmentation Algorithms using Deformable Models: A Review

Abstract Image segmentation is an important step in image processing. Major advances in the field of medical imaging have provided physicians with powerful and non-invasive techniques to probe the structure, function, and pathology of the human body. This information requires significant innovations in all aspect of image processing. Manual tracing of object boundaries generally suffers from poor reproducibility of results and it is also tedious and time consuming. Further, manual segmentations are often restricted to two dimensional (2D) slice-wise processing, which often suffers from inconsistency across segmented slices. Quantitative analysis of medical images requires reproducible, accurate, and efficient segmentation methods. By making use of medical image segmentation approaches accurate, repeatable, quantitative data must be efficiently extracted in order to support the spectrum of biomedical investigations and clinical activities, from diagnosis to radiotherapy as well as to surgery. Purpose of this review is to identify the set of methods that have been used for medical image segmentation over the past three decades, and to provide an opportunity to view the transitions that have occurred as this research area has developed. In this paper, various approaches of medical image segmentation, available algorithms for deformable models are reviewed and their advantages, disadvantages, and limitations are discussed.

Robust Multiple Image Watermarking Scheme using Discrete Cosine Transform with Multiple Descriptions

A novel oblivious and robust multiple image watermarking scheme using Multiple Descriptions (MD) and Quantization Index Modulation (QIM) of the host image is presented in this paper. Watermark embedding is done at two stages. In the first stage, Discrete Cosine Transform (DCT) of odd description of the host image is computed. The watermark image is embedded in the resulting DC coefficients. In the second stage, a copy of the watermark image is embedded in the watermarked image generated at the first stage. This enables us to achieve robustness to both local and global attacks. This algorithm is highly robust for different attacks on the watermarked image and superior in terms of Peak Signal to Noise Ratio (PSNR) and Normalized Cross correlation (NC).

Adaptive AC-Coefficient Prediction for Image Compression and Blind Watermarking

In this work, an adaptive image compression algorithm is proposed based on the prediction of AC coefficients in Discrete Cosine Transform (DCT) block during reconstruction of image. In the prediction phase, DC values of the nearest neighbour DCT blocks is utilized to predict the AC coefficients of centre block. Surrounding DC values of a DCT blocks are adaptively weighed for AC coefficients prediction. Linear programming is used to calculate the weights with respect to the image content. Results show that this method is good in terms of good Peak Signal to Noise Ratio (PSNR) and less blocking artifacts. The proposed scheme has been demonstrated through several experiments including Lena. Reconstructed image is of good quality with same compression ratio compared to the existing technique in the literature. In addition, an image watermarking algorithm is proposed using DCT AC coefficients obtained. The performance of the proposed watermarking scheme is measured in terms of PSNR and Normalized Cross Correlation (NCC). Further, this algorithm is robust for various attacks including JPEG compression on watermarked image.

STEREO MATCHING ALGORITHMS BASED ON FUZZY APPROACH

Stereo matching is the central problem of stereovision paradigm. Area-based techniques provide the dense disparity maps and hence they are preferred for stereo correspondence. Normalized cross correlation (NCC), sum of squared differences (SSD) and sum of absolute differences (SAD) are the linear correlation measures generally used in the area-based techniques for stereo matching. In this paper, similarity measure for stereo matching based on fuzzy relations is used to establish the correspondence in the presence of intensity variations in stereo images. The strength of relationship of fuzzified data of two windows in the left image and the right image of stereo image pair is determined by considering the appropriate fuzzy aggregation operators. However, these measures fail to establish correspondence of the pixels in the stereo images in the presence of occluded pixels in the corresponding windows. Another stereo matching algorithm based on fuzzy relations of fuzzy data is used for stereo matching in such regions of images. This algorithm is based on weighted normalized cross correlation (WNCC) of the intensity data in the left and the right windows of stereo image pair. The properties of the similarity measures used in these algorithms are also discussed. Experiments with various real stereo images prove the superiority of these algorithms over normalized cross correlation (NCC) under nonideal conditions.

IMAGE COMPRESSION AND WATERMARKING SCHEME USING SCALAR QUANTIZATION

This paper presents a new compression technique and image watermarking algorithm based on Contourlet Transform (CT). For image compression, an energy based quantization is used. Scalar quantization is explored for image watermarking. Double filter bank structure is used in CT. The Laplacian Pyramid (LP) is used to capture the point discontinuities, and then followed by a Directional Filter Bank (DFB) to link point discontinuities. The coefficients of down sampled low pass version of LP decomposed image are re-ordered in a pre-determined manner and prediction algorithm is used to reduce entropy (bits/pixel). In addition, the coefficients of CT are quantized based on the energy in the particular band. The superiority of proposed algorithm to JPEG is observed in terms of reduced blocking artifacts. The results are also compared with wavelet transform (WT). Superiority of CT to WT is observed when the image contains more contours. The watermark image is embedded in the low pass image of contourlet decomposition. The watermark can be extracted with minimum error. In terms of PSNR, the visual quality of the watermarked image is exceptional. The proposed algorithm is robust to many image attacks and suitable for copyright protection applications.

A Genetic Algorithm Based Oblivious Image Watermarking Scheme Using Singular Value Decomposition (SVD)

Digital Watermarking and data hiding have gained popularity in recent years as a means of protecting digital images from theft, illegal copying and unlawful reproduction. Some digital watermarking algorithms were proposed using spatial domain and transform domain techniques. The transform domain could be DFT, DCT, DWT or SVD. Several digital watermarking algorithms using Genetic Algorithms are available in the literature. In this paper, a novel optimal watermarking scheme based on Singular Value Decomposition using Genetic Algorithm (GA) is proposed. The proposed scheme is based on step size optimization using the genetic algorithm to improve the quality of watermarked image and robustness of the watermark. The proposed algorithm is more secure and robust to various attacks, viz., Low Pass Filtering, Median Filtering, JPEG Compression, Resizing, Row-Column blanking, Row-Column Copying etc. Superior experimental results were observed with the proposed scheme over a recent algorithm proposed by Chandra Mohan et al. in terms of Normalized Cross correlation(NC) and Peak Signal to Noise Ratio(PSNR).

Identification of cheaters in elections using steganography

Abstract This paper presents a new secret sharing scheme for identification of cheaters in the elections. In a proposed scheme secret image (voter ID) is embedded into cover image (Indian emblem) of same size and generates n stego images. The resulting stego images are visually good and look like a cover image. Various image quality metrics are used to assess the quality of stego images. Out of n stego images only last two stego images are required to reconstruct the secret and cover images without any loss. The proposed scheme offers high embedding capacity. Experimental results of proposed scheme shown to be superior in terms of PSNR and visual quality than existing schemes.AbstractThis paper presents a new secret sharing scheme for identification of cheaters in the elections. In a proposed scheme secret image (voter ID) is embedded into cover image (Indian emblem) of same size and generates n stego images. The resulting stego images are visually good and look like a cover image. Various image quality metrics are used to assess the quality of stego images. Out of n stego images only last two stego images are required to reconstruct the secret and cover images without any loss. The proposed scheme offers high embedding capacity. Experimental results of proposed scheme shown to be superior in terms of PSNR and visual quality than existing schemes.

HVS based Robust Image Watermarking Scheme using Slant Transform

This paper presents a robust algorithm for digital image watermarking based on Human Visual System (HVS). Watermark is embedded in the Slant Transform domain by altering the transform coefficients. The perceptibility of the watermarked image using proposed algorithm is improved over DCT based algorithm9 by embedding the watermark image in selected positions based on the HVS weightage matrix. The proposed method is robust and the watermark image can survive to many image attacks like noise, bit plane removal, cropping, histogram equalization, rotation, and sharpening. Results are compared with DCT based watermarking method and found to be superior in terms of the quality of the watermarked image and resilience to attacks. The metrics used to test the robustness of the proposed algorithm are Peak Signal to Noise Ratio (PSNR) and Normalized Cross Correlation (NCC).