Visual Information Extraction on EPHOIE (test)
99.21Mean AccuracyLayoutLMv3-Chinese
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| LayoutLMv3-ChineseModel size=base2022.04 | 99.21 | 98.99 | 100 | 99.77 | 100 | 100 | 98.82 | 99.78 | 98.31 | 97.27 | — | |
| StrucTexT2022.04 | 97.95 | 99.25 | 100 | 99.47 | 97.98 | 95.43 | 98.29 | 97.33 | 99.25 | 93.73 | — | |
| VIES2022.04 | 95.23 | 99.39 | 100 | 99.67 | 91.81 | 88.73 | 99.29 | 89.47 | 98.35 | 86.27 | — | |
| TRIE2022.04 | 93.21 | 98.79 | 100 | 99.46 | 88.64 | 85.92 | 97.94 | 84.32 | 97.02 | 80.39 | — | |
| GCN-based2022.04 | 92.55 | 98.18 | 100 | 99.52 | 88.17 | 86 | 97.39 | 80 | 94.44 | 81.82 | — | |
| GraphIE2022.04 | 90.26 | 94 | 100 | 95.84 | 82.19 | 84.44 | 93.07 | 85.33 | 94.44 | 76.19 | — | |
| BILSTM+CRF2022.04 | 89.1 | 98.51 | 100 | 98.87 | 75.86 | 72.73 | 94.04 | 84.44 | 98.18 | 69.57 | — | |
| Qwen2-VL-7BTraining Data=RIDGE + RIDGE-DS2025.04 | 80.91 | — | — | — | — | — | — | — | — | — | 89.82 | |
| Qwen2-VL-7BTraining Data=RIDGE2025.04 | 77.89 | — | — | — | — | — | — | — | — | — | 87.79 | |
| Qwen2-VL-7B+zero-shot=true, RIDGE fine-tuned=true2025.04 | 77.89 | — | — | — | — | — | — | — | — | — | 87.79 | |
| Qwen2-VL-7BTraining Data=Original2025.04 | 76.91 | — | — | — | — | — | — | — | — | — | 86.52 | |
| Qwen2-VL-7Bzero-shot=true, RIDGE fine-tuned=false2025.04 | 76.91 | — | — | — | — | — | — | — | — | — | 86.52 | |
| Monkeyzero-shot=true, RIDGE fine-tuned=false2025.04 | 22.44 | — | — | — | — | — | — | — | — | — | 30.96 | |
| LLaVA-NeXT-Mistral-7B+zero-shot=true, RIDGE fine-tuned=true2025.04 | 5.39 | — | — | — | — | — | — | — | — | — | 12.07 | |
| DocOwl-1.5-Chatzero-shot=true, RIDGE fine-tuned=false2025.04 | 4.85 | — | — | — | — | — | — | — | — | — | 8.25 | |
| LLaVA-NeXT-Mistral-7Bzero-shot=true, RIDGE fine-tuned=false2025.04 | 4.31 | — | — | — | — | — | — | — | — | — | 8.02 |