Bayer CFA image zooming using a normalized color-ratio model

A new zooming algorithm suitable for Bayer color filter array (CFA) based digital cameras is introduced. The proposed method utilizes both the spatial and the spectral correlation among the CFA inputs to achieve a high-quality, enlarged CFA image. Operating directly on the noise-free data, the zooming method avoids amplification of the impairments which are usually introduced during demosaicking. The utilization of a normalized color-ratio model in the proposed zooming scheme reduces color shifts in the enlarged CFA image and thus, the cycle completed using a CFA interpolation scheme results in full color, enlarged image pleasing for viewing.