Super-Resolution of multiple moving 3d objects with pixel-based registration

  • Authors:

  • Takuma Yamaguchi;Hiroshi Kawasaki;Ryo Furukawa;Toshihiro Nakayama

  • Affiliations:

  • Research Center, The Nippon Signal Co., LTD, Saitama, Japan;Department of Information and Computer Sciences, Saitama University, Saitama, Japan;Department of Computer Science, Hiroshima City University, Hiroshima, Japan;Research Center, The Nippon Signal Co., LTD, Saitama, Japan

  • Venue:
  • ACCV'09 Proceedings of the 9th Asian conference on Computer Vision - Volume Part III
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we propose a super-resolution technique for multiple independently moving 3D objects from a single camera, the camera is also allowed to move freely Previous techniques were mostly focused on planar objects, and it was difficult to realize super-resolution for the scene containing 3D objects, due to significant appearance changes caused by objects' motion and viewpoint changes In this paper, we propose a new technique which can solve the above mentioned problems by applying pixel-based registration instead of planar based registration, which are commonly used in the previous 3D super-resolution techniques Since the technique is pixel-wise and is not required to divide the scene into planar patches, it can be applied to images containing objects with complex shapes or non-rigid objects, to which applying planar approximation is difficult.