In-Image Machine Translation Evaluation on Vectra-Bench zh→pt 1.0
0.773Pearson Correlation (r)Vectra
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VectraProtocol=No-Reference (Direct Assessment)2026.01 | 0.773 | 0.675 | |
| FIDProtocol=Reference-based (Top-Ranked Candidate as Reference)2026.01 | 0.695 | 0.599 | |
| HCIIT-JudgeProtocol=No-Reference (Direct Assessment)2026.01 | 0.647 | 0.544 | |
| PSNRProtocol=Reference-based (Top-Ranked Candidate as Reference)2026.01 | 0.59 | 0.485 | |
| SSIMProtocol=Reference-based (Top-Ranked Candidate as Reference)2026.01 | 0.586 | 0.476 | |
| LPIPSProtocol=Reference-based (Top-Ranked Candidate as Reference)2026.01 | 0.368 | 0.296 | |
| SAMScoreProtocol=No-Reference (Direct Assessment)2026.01 | 0.364 | 0.257 | |
| E-Commerce VQAProtocol=No-Reference (Direct Assessment)2026.01 | 0.044 | 0.025 |