Temporal Edges: The Detection of Motion and the Computation of Optical Flow
Abstract
A new method for the detection of motion and the computation of optical flow is presented. In the first step of the calculation the intensity history at each pixel is convolved with the second derivative in time of a temporal Gaussian smoothing function. The zero crossings in a single frame of the resulting function indicate the positions of moving edges. Spatial and temporal derivatives of the function at the zero-crossing locations are then used to compute the component of the flow that is normal to the zero-crossing contours. Both the detection of motion and the computation of the normal velocity are insensitive to slow temporal and spatial changes in the image intensity that are caused by illumination effects rather than motion. A framework in which to relate the present work to a number of gradient based flow measurement techniques is also presented.
Cite
Text
Duncan and Chou. "Temporal Edges: The Detection of Motion and the Computation of Optical Flow." IEEE/CVF International Conference on Computer Vision, 1988. doi:10.1109/CCV.1988.590014Markdown
[Duncan and Chou. "Temporal Edges: The Detection of Motion and the Computation of Optical Flow." IEEE/CVF International Conference on Computer Vision, 1988.](https://mlanthology.org/iccv/1988/duncan1988iccv-temporal/) doi:10.1109/CCV.1988.590014BibTeX
@inproceedings{duncan1988iccv-temporal,
title = {{Temporal Edges: The Detection of Motion and the Computation of Optical Flow}},
author = {Duncan, James H. and Chou, Tsai-Chia},
booktitle = {IEEE/CVF International Conference on Computer Vision},
year = {1988},
pages = {374-382},
doi = {10.1109/CCV.1988.590014},
url = {https://mlanthology.org/iccv/1988/duncan1988iccv-temporal/}
}