Papers by Karan Jariwala
VQD: Visual Query Detection In Natural Scenes (N19-1)
Copied to clipboard
| Challenge: | Existing visual referring expression recognition tasks have multiple annotation problems and language bias problems. |
| Approach: | They propose a visual grounding task called Visual Query Detection . they evaluate the first algorithms on visual referring expression datasets and VQDv1 datasets . |
| Outcome: | The proposed algorithms are compared with existing visual referring expression comprehension datasets and the new VQDv1 dataset. |