Title :
Compressive oscillatory patterns sensing in natural images via iterative wave atom thresholding
Abstract :
This paper focuses on the problem of oscillatory pattern reconstruction in natural images from the perspective of compressive sensing. An iterative wave atom thresholding method for the recovery of oscillatory signals is presented. This allows a better reconstruction of natural images from highly incomplete measurements. Numerical experiments on real scenes demonstrate that the method capture the structures of complex oscillatory patterns.
Keywords :
image reconstruction; image representation; iterative methods; compressive oscillatory pattern sensing; compressive sensing; iterative wave atom thresholding; natural image; pattern reconstruction; Atomic measurements; Compressed sensing; Image reconstruction; Iterative methods; PSNR; Wavelet transforms;
Conference_Titel :
Signal Processing (ICSP), 2010 IEEE 10th International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4244-5897-4
DOI :
10.1109/ICOSP.2010.5655915