Image compression using wavelet based compressed sensing and vector quantization

Sparse signal representations and compressed sensing have found use in a large number of applications including image compression. Compressed sensing exploits the sparsity of naturally occurring images to reduce the volume of the data by using less number of measurements. Inspired by this, we propose a new framework for image compression that combines compressed sensing theory with wavelet and vector quantization. Wavelet transform is used to sparsify the input image while measurement vectors generated from the sparse vectors are transmitted using vector quantization. Simulation experiments are carried out to analyze the effects of various parameters on the image reconstruction quality. Results obtained have been found to be quite promising.