Abstract | ||
---|---|---|
We propose an adaptive robust estimation algorithm for filtering motion vectors. We first extract motion vectors from consecutive images by using size-variable block matching and then apply the extracted motion vectors to adaptive robust estimation to filter them. The proposed robust estimation defines a sigmoid weight function, and eliminates outliers by gradually tuning the sigmoid function to the hard limit as the error between model parameters and input data is minimized. |
Year | DOI | Venue |
---|---|---|
2002 | 10.1007/3-540-45813-1_74 | Rough Sets and Current Trends in Computing |
Keywords | Field | DocType |
input data,consecutive image,sigmoid function,adaptive robust estimation algorithm,filtering motion vectors,sigmoid weight function,extract motion vector,motion vector,robust estimation,adaptive robust estimation,hard limit,proposed robust estimation,robust estimator,weight function | Mathematical optimization,Weight function,Quarter-pixel motion,Computer science,Filter (signal processing),Outlier,Algorithm,Adaptive algorithm,Motion estimation,Sigmoid function,Distributed computing,Motion vector | Conference |
Volume | ISSN | ISBN |
2475 | 0302-9743 | 3-540-44274-X |
Citations | PageRank | References |
0 | 0.34 | 4 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Seok-Woo Jang | 1 | 55 | 12.72 |
Essam A. El-Kwae | 2 | 111 | 11.06 |
Hyung-Il Choi | 3 | 138 | 26.28 |