Fast Directional Dictionary Interpolation Super Resolution Algorithm in Medium and Low Magnification

• Main Author: WANG Zhongfei, WANG Biao
• Format: Article
• Language: zho
• Published: Journal of Computer Engineering and Applications Beijing Co., Ltd., Science Press 2020-02-01
• Series: Jisuanji kexue yu tansuo
• Subjects: super resolution; medium and low magnification; texture; interpolation
• Online Access: http://fcst.ceaj.org/CN/abstract/abstract2123.shtml

In a real-time medium and low magnification super resolution display system, a fast medium and low magnification super resolution algorithm using all directional dictionary and fitting interpolation is proposed to improve the reconstruction image performance and reduce the computational complexity. For every interpolated pixel in high resolution image, first of all, the main texture direction of the corresponding interpolation interval in low resolution image is obtained through the pyramid texture dictionary. Then, the interpolation direction is modified based on the interpolated pixel position. Finally, the interpolation result is obtained by the unidirectional fitting interpolation. Simulation results show that compared with SREO (super resolution using edge orientation based on linear mapping), the proposed algorithm improves PSNR (peak signal to noise ratio) and SSIM (structural similarity) by 0.73 dB and 0.04 averagely and respectively. Moreover, the average computation time is reduced about 40% in the proposed algorithm.