Color demosaicing via directional linear minimum mean square-error estimation

$49

Description

This project presented a new color demosaicking technique of LMMSE directional filtering of the green-red and green-blue PDS signals. The missing green samples are estimated from the filtered PDS in both horizontal and vertical directions, and the two estimates are optimally fused. The resulting green channel is then used to guide the estimation of the missing red and blue samples. 

The experiments showed that the proposed color demosaicking algorithm significantly outperformed the current state of the art demosaicing methods both in PSNR measure and visual quality. Furthermore, the proposed algorithm is non-iterative, fast, and easy to implement.

 

ref :

Zhang, L., & Wu, X. (2005). Color demosaicking via directional linear minimum mean square-error estimation. IEEE Transactions on Image Processing, 14(12), 2167-2178.

A joint demosaicing-zooming scheme for single sensor digital color cameras

Multiscale KSVD Denoising

 

Reviews

There are no reviews yet.

Be the first to review “Color demosaicing via directional linear minimum mean square-error estimation”