Pointing Game Accuracy: 55.91
Papers with Code · CC BY-SA 4.0
Phrase Grounding
Given an image and a corresponding caption, the Phrase Grounding task aims to ground each entity mentioned by a noun phrase in the caption to a region in the image.
<span class="description-source">Source: Phrase Grounding by Soft-Label Chain Conditional Random Field </span>
논문
R@1: 41.77R@10: 70.77R@5: 64.52
Pointing Game Accuracy: 55.16
R@1: 87.1R@10: 97.4R@5: 96.1
Accuracy: 28.91
R@1: 84.1
R@1: 70.4R@10: 86.31R@5: 84.49
R@1: 69.69R@10: 86.35R@5: 84.22
R@1: 78.73
R@1: 87.7
R@1: 87.1R@10: 98.1R@5: 96.9
R@1: 48.38
R@1: 76.74
R@1: 43.89R@10: 68.66R@5: 64.46
R@1: 84.3R@10: 95.8R@5: 93.9
R@1: 27.8R@10: 62.9
R@1: 74.69
R@1: 73.3