Research Article
Image Retrieval Based on a Multi-Integration Features Model
Table 4
The performance of the proposed method according to different distance or similarity metrics.
| Dataset | Performance (%) | Distance or similarity metrics | Canberra | x 2 statistics | L 1 | Histogram intersection | Cos | Chebyshev |
| Corel-10K | Precision | 44.63 | 45.42 | 52.96 | 24.03 | 48.61 | 35.16 | Recall | 5.35 | 5.45 | 6.36 | 2.88 | 5.83 | 4.22 |
| GHIM-10K | Precision | 50.33 | 51.4 | 56.48 | 31.01 | 54.87 | 39.98 | Recall | 1.21 | 1.23 | 1.36 | 0.74 | 1.32 | 0.96 |
|
|