Search

An approach for the combined image authentication and compression of color images by making use of digital watermarking and data hiding framework is presented. The digital watermark is comprised of two components that include a soft-authenticator watermark for authentication and tamper assessment of the given image, and a chrominance watermark employed to improve the efficiency of compression.

A study of the performance analysis and design of robust watermark embedding strategies in the presence of block-based lossy compression is presented. A novel hybrid watermarking algorithm that exploits the best of robust digital watermarking techniques is developed for greater resilience to JPEG compression through analytical comparisons.

Hardware implementation aspects of the digital watermarking problem are investigated and discussed.

Communication theory principles of diversity and channel estimation are used with the aim of strengthening the robustness of digital watermarking procedures.

Robust classification of blurred imagery using two novel approaches that involve data fusion is discussed. The problem addressed is called blind image fusion and techniques used the nonnegativity and support constraints recursive inverse filtering algorithm for blind image restoration and the Markov random field-based fusion method for classification. Simulations results on synthetic and real photographic data show the potential the approaches have.

A general framework based on Lyapunov's direct technique for studying the behavior of nonlinear blind deconvolution algorithms was analyzed. The framework was applied to the blind deconvolution scheme such as the nonnegativity and support constraints recursive inverse filtering (NAS-RIF) algorithm for blind image restoration. The method was effective in generating adequate conditions for the distinct global convergence of an iterative blind deconvolution algorithm and offered a unique solution to the NAS-RIF algorithm.

A study was conducted on a novel blind deconvolution approach for the restoration of linearly degraded images using recursive filtering. The method is applicable to situations in which the scene consists of a finite support object against a uniformly black, grey or white background. The findings indicate that the new technique is more reliable and has a faster convergence than existing nonparametric finite support blind deconvolution methods.

Research describing the classification of blurred images is presented. The concept of blind image fusion is developed and algorithms derived which are believed to enhance the process of image restoration.