Bootstrap-based method for occlusion removal in case of multiple sensors

  • Authors:
  • Maodong Jiang;Layachi Bentabet

  • Affiliations:
  • University of Sherbrooke, Sherbrooke, Canada;Bishop's University, Sherbrooke, Canada

  • Venue:
  • SPPRA '07 Proceedings of the Fourth IASTED International Conference on Signal Processing, Pattern Recognition, and Applications
  • Year:
  • 2007

Quantified Score

Hi-index 0.01

Visualization

Abstract

In this paper, a new method for occlusion removal, in the presence of multiple images describing the same scene, is proposed. Our approach integrates the information from the non occluded sensors into the occluded ones. A data fusion step provides a global clustering of the scene, where each cluster is represented by a Dirichlet distribution. This information is re-expressed in terms of the characteristics of the occluded sensor using the bootstrap technique. In order to preserve local image characteristics, we propose to achieve the restoration using an effective combination of exemplar-based inpainting and moving blocks bootstrap. The effectiveness of our method is demonstrated on synthetic and radar and SPOT images.