Multimedia applications facilitate the users to integrate and manipulate data from various sources such as video, images, animation, graphics, audio and text in one single hardware platform. With the exponential growth of multimedia applications, the potential to index, store and retrieve multimedia data in an efficient way has become a challenge for researchers. Digital images have become an important element of multimedia applications. Images help to express information in an art form and often account for large storage space. As a result, digital image compression has become a spotlight in our digital field. It is also essential that the image quality is not deteriorated on compression of images. In the recent years, various image compression algorithms and techniques have been developed. Each application has an image compression algorithm with intent to increase either the compression ratio or the image quality of the reconstructed image. Numerous applications which include compression of images are now newfangled by the use of sparse representation and compressive sensing. Many authors have refined numerous algorithms for compressive sensing that customize discrete wavelet transform (DWT) with single level of decomposition along with orthogonal wavelets to obtain good reconstruction of an image. This paper contributes to show an improvement in the compression ratio by two level decomposition of an image through two orthogonal wavelets. Daubechies5 (db5) is used in the first level of decomposition and Daubechies8 (db8) is employed in the second level of decomposition. By the combination of db5 and db8, the best coefficients for image compression are obtained having exclusive properties of energy compaction, maximal vanishing moments, compact support, symmetry and regularity. The measurement matrix used is Gaussian random matrix and the image is reconstructed by orthogonal matching pursuit (OMP) reconstruction algorithm. The image quality metrics such as compression ratio (CR), mean square error (MSE), peak signal to noise ratio (PSNR) and structural similarity index (SSIM) are compared for the proposed work with the existing compressive sensing technique in literature. Various images are taken for analysis and the proposed work shows improvement in compression ratio by 20-35%.

integration, orthogonal wavelets.