Research Article
Deep Visual Semantic Embedding with Text Data Augmentation and Word Embedding Initialization
Table 5
Experimental results with text data augmentation on MS-COCO.
| | Model | Feature (s) | Image count | Image retrieval | | R@1 | R@5 | R@10 | Med r | R@1 | R@5 | R@10 | Med r |
| | VSA | R–CNN + BRNN | 38.4 | 69.9 | 80.5 | 1 | 27.4 | 60.2 | 74.8 | 3 | | VSE++ | VGG + GRU + HNM | 43.6 | 74.8 | 84.6 | 2 | 33.7 | 68.8 | 81.0 | 3 | | Ours | Aug | 45.1 | 75.8 | 85.3 | 2 | 33.8 | 67.4 | 80.2 | 3 |
|
|
