Research Article
Lower Order Krawtchouk Moment-Based Feature-Set for Hand Gesture Recognition
Table 5
Comparison of results for Jochen-Triesch’s dataset.
| Reference | Feature extraction method | Recognition accuracy |
| [14] | Elastic graph matching | 92.9% |
| [26] | Modified census transform | 89.9% |
| [27] | Weighted Eigen-space size functions | 85.1% |
| [28] | Tchebichef (till 9th order), Hu, and Geometric features | 84.63% (dark background) 95.55% (light background) |
| Proposed method | Krawtchouk features till 3rd order (45 features) | 84.9% |
|
|