Fast 3-D interpretation from monocular image sequences on large motion fields

Authors:
Jong-Sung Kim;Ki-Sang Hong
Affiliations:
Division of Electrical and Computer Engineering, POSTECH, Pohang, Korea;Division of Electrical and Computer Engineering, POSTECH, Pohang, Korea
Venue:
ACCV'07 Proceedings of the 8th Asian conference on Computer vision - Volume Part II
Year:
2007

Citing 8
Cited 0

An Investigation of Smoothness Constraints for the Estimation of Displacement Vector Fields from Image Sequences

IEEE Transactions on Pattern Analysis and Machine Intelligence
Computational analysis of visual motion

Computational analysis of visual motion
Model-Based Brightness Constraints: On Direct Estimation of Structure and Motion

IEEE Transactions on Pattern Analysis and Machine Intelligence
Direct Recovery of Planar-Parallax from Multiple Frames

IEEE Transactions on Pattern Analysis and Machine Intelligence
Dense 3D Interpretation of Image Sequences: A Variational Approach Using Anisotropic Diffusion

ICIAP '03 Proceedings of the 12th International Conference on Image Analysis and Processing
Dense 3D Interpretation of Image Sequences: A Variational Approach Using Anisotropic Diffusion

ICIAP '03 Proceedings of the 12th International Conference on Image Analysis and Processing
Three-Dimensional Scene Flow

IEEE Transactions on Pattern Analysis and Machine Intelligence
Efficient and reliable schemes for nonlinear diffusion filtering

IEEE Transactions on Image Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper proposes a fast method for dense 3-D interpretation to directly estimate a dense map of relative depth and motion from a monocular sequence of images on large motion fields. The Nagel-Enkelmann technique is employed in the variational formulation of the problem. Diffusion-reaction equations are derived from the formulation so as to approximate the dense map on large motion fields and realize an anisotropic diffusion to preserve the discontinuities of the map. By combining the ideas of implicit schemes and multigrid methods, we present a new implicit multigrid block Gauss-Seidel relaxation scheme, which dramatically reduces the computation time for solving the largescale linear system of diffusion-reaction equations. Using our method, we perform fast 3-D interpretation of image sequences with large motion fields. The efficiency and effectiveness of our method are experimentally verified with synthetic and real image sequences.