Research Article
Triplet Deep Hashing with Joint Supervised Loss Based on Deep Neural Networks
Table 4
Dataset settings used in our experiment.
| Dataset | Total | Training (randomly) | Testing (randomly) | Labels |
| CIFAR-10 | 60,000 | 5,000 (500 ∗ 10) | 1,000 (100 ∗ 10) | 10 | NUS-WIDE | 269,648 | 10,500 (500 ∗ 21) | 2,100 (100 ∗ 21) | 21 | MS-COCO | 82,783 | 10,000 | 5,000 | 80 |
|
|