2 March 1994 Computation of the depth from motion using a massively parallel neural network approach
Author Affiliations +
Abstract
As many early vision tasks the computation of depth-from-motion is an ill-posed problem but very useful in computer vision and rotor craft navigation. The collective computation capabilities of highly parallel neural networks provides new powerful techniques for optimization problems in high dimensional spaces. This paper reports an investigation of computation of depth from motion. As this problem is formulated as minimizing a cost or energy function, a massively parallel neural network approach is used for solving this problem by regularization techniques. This approach presents some similarities with biological visual systems. The neural solution developed here is a direct method avoiding the explicit optical flow estimation. We perform an evaluation on both synthetic and real world image sequence.
© (1994) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jean-Luc Sune, Pierre Puget, Roger A. Samy, "Computation of the depth from motion using a massively parallel neural network approach", Proc. SPIE 2243, Applications of Artificial Neural Networks V, (2 March 1994); doi: 10.1117/12.169978; https://doi.org/10.1117/12.169978
PROCEEDINGS
9 PAGES


SHARE
RELATED CONTENT

Depth Cues From Optical Flow
Proceedings of SPIE (March 01 1990)
Depth determination using complex logarithmic mapping
Proceedings of SPIE (February 01 1991)
Model for spatial and chromatic vision
Proceedings of SPIE (March 01 1992)
Door detection in images based on learning by components
Proceedings of SPIE (October 05 2001)
Visual tracking without feedback control
Proceedings of SPIE (August 06 1993)

Back to Top