An Effective Demosaicking and Zooming Algorithm for CFA Images

Demosaicking interpolation and digital zooming are two significant techniques for the quality of enlarged full-color image in resource-constrained single chip devices. This paper proposes an effective demosaicking and zooming algorithm by referencing to the neighboring information. In addition to considering one direction neighboring information to estimate the missing value, the EDZ algorithm develops expressions to join the horizontal, vertical, and diagonal neighboring information to improve the accuracy of the estimated value. Experiments performed on the EDZ algorithm and previous schemes show that the EDZ algorithm achieves better visual perception and the PSNR measurement in average.