Image Captioning 벤치마크
Image Captioning on nocaps-XD out-of-domain
CIDEr
- 2020-09-28 — Microsoft Cognitive Services team: CIDEr 95.5
- 2022-05-27 — GIT2: CIDEr 122.27
| Rank | Model | CIDEr | B1 | B2 | B3 | B4 | ROUGE-L | METEOR | SPICE | Paper | Code | Year |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | GIT2 | 122.27 | 86.28 | 71.15 | 52.36 | 30.15 | 60.91 | 30.15 | 15.62 | GIT: A Generative Image-to-text Transformer for Vision and Language | microsoft/GenerativeImage2Text | 2022 |
| 2 | GIT | 122.04 | 85.99 | 71.28 | 52.66 | 30.04 | 60.96 | 30.45 | 15.7 | GIT: A Generative Image-to-text Transformer for Vision and Language | microsoft/GenerativeImage2Text | 2022 |
| 3 | Microsoft Cognitive Services team | 95.5 | 79.44 | 61.15 | 41.03 | 21.79 | 55.49 | 26.56 | 12.66 | VIVO: Visual Vocabulary Pre-Training for Novel Object Captioning | 2020 | |
| 4 | Human | 91.62 | 74.84 | 53.9 | 33.51 | 16.6 | 51.5 | 26.83 | 14.21 | |||
| 5 | VLAF2 | 90.34 | 79.59 | 61.04 | 40.09 | 19.61 | 54.86 | 26.14 | 13.11 | |||
| 6 | icp2ssi1_coco_si_0.02_5_test | 85.28 | 75.59 | 56.71 | 35.63 | 17.72 | 51.92 | 23.77 | 11.28 | |||
| 7 | test_cbs2 | 77.94 | 74.5 | 53.63 | 30.91 | 13.41 | 49.66 | 23.47 | 11.07 | |||
| 8 | UpDown + ELMo + CBS | 66.67 | 71.57 | 48.58 | 25.77 | 9.68 | 47.13 | 20.88 | 9.74 | |||
| 9 | Neural Baby Talk + CBS | 58.48 | 65.98 | 43.2 | 21.16 | 7.5 | 44.47 | 19.04 | 8.77 | |||
| 10 | Neural Baby Talk | 48.73 | 64.45 | 42.8 | 21.48 | 7.92 | 44.11 | 18.31 | 8.2 | |||
| 11 | UpDown | 30.09 | 66.54 | 44.28 | 24.23 | 10.17 | 44.84 | 18.29 | 8.08 |