1. INTRODUCTION
CFA images are familiar in most digital cameras, mobile
phone and photograph. In each pixel of CFA image, there is
only one color. Green, red and blue pixels are put and
arranged alternately. There are 50% of green pixels, 25% of
red pixels, and 25% of blue pixels. To formulate a full color
image from CFA image, the two missing color values at
each pixel are usually estimated from CFA neighboring
pixels. The previous methods are performed by using basic
element of the image. Normalized color-ratio model was
used [1]. It utilized both spatial and spectral correlation to
reduce color shifts in the enlarged CFA. Bayer pattern was
considered by using a color-ratio model and an edge-sensing
mechanism to produce naturally color sharp enlarged output
[2]. Spectrum model was used to preserve spectrum