A novel Ray-space based view generation algorithm via Radon transform

  • Authors:
  • Lingfeng Xu;Ling Hou;Oscar C. Au;Wenxiu Sun;Xingyu Zhang;Yuanfang Guo

  • Affiliations:
  • Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong, China;Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong, China;Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong, China;Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong, China;Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong, China;Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong, China

  • Venue:
  • 3D Research
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Ray-space interpolation is one of the key technologies to generate virtual viewpoint images, typically in epipolar plane images (EPI), to realize free viewpoint television (FTV). In this paper, a novel Radon transform based ray-space interpolation algorithm (RTI) is proposed to generate virtual viewpoint images in the EPI. In the proposed RTI, feature points of each EPI are first extracted to form the corresponding feature epipolar plane image (FEPI). Radon transform is then applied to each FEPI to detect the candidate interpolation direction set. Then corresponding pixels in neighboring real view rows for each pixel to be interpolated are found by some block matching based interpolation method, in which the smoothness property of the disparity field and the correlation among neighboring EPIs are explored. Possible occlusion regions are processed by the proposed one-sided interpolation. Finally, to solve the problem that the cameras are not equally spaced, a novel ray-space based spacing correction algorithm is proposed to correct the EPI such that pixels corresponding to the same scene point will form a straight line. Experimental results suggest that the proposed RTI and the spacing correction algorithm can achieve good performance. Moreover, compare with traditional methods, our proposed algorithm can generate good interpolation result even if apriori depth range is not known.