Name
Abstract | ||
|---|---|---|
This paper introduces a low-level visual feature which is an extension of the maximally stable extremal region (MSER) to color, applying for sports video genre categorization. The extension to color is done by detecting and describing the features based on opponent color space instead of gray-level in an image. The proposed feature is invariant not only to scale, rotation and affine transform, but also to light intensity change and shift (illumination). We compare our algorithm on the classification average accuracy to the state-of-art local invariant visual features including the original MSER, the original scale invariant feature transform (SIFT) and color-based SIFT. The experiment result illustrates that the average accuracy based on the proposed algorithm is 7.09%, 6.49% and 21.4% higher than the other three algorithms respectively. |
Year | DOI | Venue |
|---|---|---|
2012 | 10.1109/CCIS.2012.6664364 | Proceedings - 2012 IEEE 2nd International Conference on Cloud Computing and Intelligence Systems, IEEE CCIS 2012 |
Keywords | DocType | Volume |
opponent color space,video signal processing,scene classification,scale invariant feature transform,sport,affine transform,genre categorization,color based maximally stable extremal region,mser,transforms,sports video genre categorization,image colour analysis,rotation transform,sift | Conference | 1 |
Issue | ISSN | ISBN |
null | null | 978-1-4673-1855-6 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Nan Zhao | 1 | 4 | 1.75 |
| Yuan DONG | 2 | 139 | 25.66 |
| Jiwei Zhang | 3 | 15 | 3.77 |
| Xiaofu Chang | 4 | 12 | 1.99 |