"Unified approach for Image Encryption Error Detection & Correction and its Low Power Architecture Implementation"N.Venkateshwaran Abhishek Gopal N.K.Bharath

WAran Research FoundaTion

AbstractImage encryption and image encoding for error detection and correction are very vital issues to ensure security and reliability during image transmission. Images are characterized by pixels and are often corrupted by noise [1]. Any change in the pixel values during transmission is due to the presence of noise. It is hence necessary to correct the faulty pixels at the receiving end. This paper presents a novel approach unifying both the encryption , error detection and correction processes. A low power architecture is suggested for implementing this methodology taking into account the sparsity and redundancy present in the images.

Matrix algorithm encoding techniques[1] are employed for unifying both the processes mentioned above. Generally partitioned matrix algorithm encoding techniques are used in a supercomputing environment for error detection and correction in applications execution involving matrix operations[1]. The image matrix is partitioned into sub-matrices. The algorithm-level-encoding technique is applied to each of the sub-matrices [1]. A suitable kernel matrix is randomly chosen and this is partitioned corresponding to the image matrix sub matrices. For each of the sub matrices belonging to both the image and the kernel the column check sum and row checksum elements are calculated. Subsequently, the Bose, Ray-Chaudhuri, Hocquenghem coding (BCH) encoding is done to only the row and column elements of the sub-matrices.It is to be noted that BCH encoding is not applied to all the pixels of the image matrix, there by reducing computation and hence power[2]. Also Randomly Sparse Matrix Multiplier [3] (RSMATMUL).RSMATMUL is employed to further reduce the dynamic power.

Referneces

Stefanidis V.K and Margaritis K.G, “Algorithm Based Fault Tolerance: Review and Experimental Study”, ICNAAM 2004.

Towards a Unified theory for Image Encryption, Encoding and Compression. Thesis submitted by Abhishek Gopal to Waran Research Foundation (WARFT), 2005.

Venkateswaran Nagarajan, Abhishek Gopal, Arun Kumar N.R and Shiv Ram Shankar Ram,Low Power Sparse Matrix Solvers for DSP Applications,Under Review in IEEE Trans on Circuits and Systems .