Wavelet Based Sparse Image Reconstruction Using Compressed Sensing Algorithm And Vector Quantization

Ordinary images, as well as most natural and manmade signals, are compressible and can, therefore, be well represented in a domain in which the signal is sparse. Compressed sensing (CS) uses a less number of linearly projected measurements to exploit the sparsity of naturally occurring images to reduce the volume of the data. Inspired by recent theoretical advances in compressed sensing, we propose a new framework for image compression that combines the compressed sensing theory with wavelet and vector quantization. Wavelet transform is used to obtain a sparse representation of the original image while vector quantization is employed for transmission of the measurement vectors generated from the sparse vectors. Results obtained have been found to be quite promising. Keyword: Image Compression, Compressed Sensing, Wavelet, Sparse, Vector Quantization.