Video Super-resolution Reconstruction based on Loca-adaptation Interpolation Wavelet Transform IBP Algorithm

Yuanyuan WuXiaohai He, Yanyue Sun, Weilong Chen

College of Electronics and Information Engineering, Sichuan University, Chengdu, China

Abstract

For the video super-resolution reconstruction based on iterative back projection (IBP) algorithm, an improved method by raising the initial guess for improving reconstruction quality is proposed. Firstly a better initial guess is performed with wavelet-local adaptation interpolation. Then Combined with segmentation reconstruction model, a super-resolution video is reconstructed from a low-resolution video associated with the promotion of spatial resolution. At last the experimental results show that the reconstructed super-resolution video has better visual effect.



Publication

Wu, Yuanyuan, Xiaohai He, Yanyue Sun, Weilong Chen. "Video super-resolution reconstruction based on local-adaptation interpolation wavelet transform IBP algorithm." Journal of Sichuan University, 48, no. 2 (2011): 349-355. [PDF] (In Chinese)


Results

    IBP with initial image reconstructed by nearest interpolation method.
    IBP with initial image reconstructed by our wavelet local-adaptation interpolation method. PSNR curves with different initial image and different number of frames.
    From top-to-down, the four curves are our improved IBP algorithm, IBP algorithm with bicubic interpolation,
    IBP algorithm with bilinear interpolation, IBP algorithm with nearest interpolation.


Reference

[1] SUN Yan-Yue, HE Xiao-Hai, CHEN Wei-Long. Image Super-resolution Reconstruction of Wavelet Local-adaptation Interpolation. Computer Engineering, 2010, 36(13): 183-185.


Last updated by Wu, Y. on January 13, 2014