This paper focuses on color image zooming performed on the Bayer pattern, a color filter array widely used in digital still cameras. The proposed method firstly enlarges the Bayer image and secondly interpolates the missing color elements of the enlarged Bayer image by an edge-sensing interpolation scheme combined with a color difference model. This enhances structural content preservation and results in excellent color fidelity of the enlarged image. The proposed edge-sensing zooming algorithm preserves the structural content of the Bayer data and produces a sharp, enlarged output. Used cascades of the proposed zooming method and efficient color filter array interpolations produce images that are naturally colored and pleasurable for viewing. This also results in excellent results in terms of objective image quality measure.
R. Lukac, K.N. Plataniotis, B. Smolka, D. Hatzinakos, "A New Color Image Zooming Technique for Digital Still Cameras" in Proc. IS&T CGIV 2004 Second European Conf. on Colour in Graphics, Imaging, and Vision, 2004, pp 515 - 519, https://doi.org/10.2352/CGIV.2004.2.1.art00103