Experimental Study of Intermediate Filters applied to Robust Optical Flow Estimation in Digital Images

By : Vanel Lazcano

Page No: 245-268

Abstract: 
Optical flow (OF) is the motion of pixels between two consecutive images. Variational models are used to estimate the OF. The flow that minimizes this model is the OF of the image sequence. The minimization procedure performs a coarse-to-fine approach. In each iteration, the estimated OF is processed with an intermediate filter (IF) to eliminate noise. We considered a robust OF model, and we tested it with different IF to improve OF estimation. Using the best IF and best parameters, we have evaluated the model in state-of-the-art dataset obtaining position 113 of 192 methods. Our approach outperforms similar techniques, such that SPARSEFLOW.

Author :
Vanel Lazcano
Universidad Mayor, Núcleo de Matemática, Físicay Estadística.
 

DOI: https://doi.org/10.32381/JCISS.2019.44.1-4.13