Image Captioning on MSCOCO (randomly sampled 100 images)
4.67DescriptivenessBLIP-SMILE
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| BLIP-SMILEoptimization=SMILE2023.06 | 4.67 | 4.05 | 4.34 | 4.75 | |
| CapEnrich2023.06 | 3.89 | 4.39 | 4.12 | 4.36 | |
| Human2023.06 | 3.53 | 4.53 | 3.97 | 4.87 | |
| BLIP2023.06 | 3.41 | 4.57 | 3.91 | 4.91 |