Regularization Parameter Determination for Optical Flow Estimation using L-curve


The KIPS Transactions:PartB , Vol. 14, No. 4, pp. 241-248, Aug. 2007
10.3745/KIPSTB.2007.14.4.241,   PDF Download:

Abstract

An L-curve corner detection method is proposed for the determination of the regularization parameter in optical flow estimation. The method locates the positive peak whose curvature difference from the just right-hand negative valley is the maximum in the curvature plot of the L-curve, while the existing curvature-method simply finds the maximum in the plot. Experimental results show that RMSE of the estimated optical flow is greater only by 0.02 pixels-per-frame than the least in the average sense. The proposed method is also compared with an existing curvature-method and the adaptive pruning method, resulting in the optical flow estimation closest to the least RMSE.


Statistics
Show / Hide Statistics

Statistics (Cumulative Counts from September 1st, 2017)
Multiple requests among the same browser session are counted as one view.
If you mouse over a chart, the values of data points will be shown.


Cite this article
[IEEE Style]
J. D. Kim and J. W. Kim, "Regularization Parameter Determination for Optical Flow Estimation using L-curve," The KIPS Transactions:PartB , vol. 14, no. 4, pp. 241-248, 2007. DOI: 10.3745/KIPSTB.2007.14.4.241.

[ACM Style]
Jong Dae Kim and Jong Won Kim. 2007. Regularization Parameter Determination for Optical Flow Estimation using L-curve. The KIPS Transactions:PartB , 14, 4, (2007), 241-248. DOI: 10.3745/KIPSTB.2007.14.4.241.