Compressive measurements with integrated sensing, compression and data processing — Initial study

In this work, modifications to the paradigm of compression using the concept of compressive sensing was proposed. Procedure of detail estimation based on image coarse approximation from linear cosine measurements was optimized. These details were used to determine the most advantageous locations of fixed noiselet measurements. Adaptively optimized calculation of sparsity pattern over significant singularities enhanced noiselet measurement efficiency. Experimentally verified effectiveness of the proposed sensed compression method has shown significant improvement in compression efficiency compared to reference methods.