ASR-OCR Alignment on Technical Presentation Slide Alignment Dataset 96 samples 1.0
99.8Avg Precision (Sc)Fuzzy
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| FuzzyThreshold Setting=T-1, Textual region threshold=0.8, Visual region threshold=0.62026.01 | 99.8 | 66.3 | 25.4 | 100 | 67.3 | 24.6 | |
| FuzzyThreshold Setting=T-2, Textual region threshold=0.7, Visual region threshold=0.62026.01 | 99.8 | 60.3 | 30.7 | 100 | 61.7 | 29.5 | |
| FuzzyThreshold Setting=T-3, Textual region threshold=0.6, Visual region threshold=0.62026.01 | 98.9 | 52.5 | 36.4 | 98.9 | 52.6 | 36 | |
| s-BERTThreshold Setting=T-1, Textual region threshold=0.8, Visual region threshold=0.62026.01 | 98.3 | 57.8 | 33.5 | 99.3 | 62.4 | 29.9 | |
| s-BERTThreshold Setting=T-2, Textual region threshold=0.7, Visual region threshold=0.62026.01 | 94.3 | 46 | 40.3 | 96.6 | 50.7 | 37.3 | |
| s-BERTThreshold Setting=T-3, Textual region threshold=0.6, Visual region threshold=0.62026.01 | 86.8 | 37.2 | 42.9 | 90.2 | 39.1 | 43 | |
| Qwen2.52026.01 | 84.3 | 49 | 34.2 | 84.5 | 49.9 | 32.3 | |
| t52026.01 | 68.7 | 27.2 | 42 | 68 | 31.4 | 40 | |
| Sci-BERTThreshold Setting=T-1, Textual region threshold=0.8, Visual region threshold=0.62026.01 | 52.4 | 35.6 | 28 | 63.6 | 47.4 | 27.9 | |
| SPECTERThreshold Setting=T-1, Textual region threshold=0.8, Visual region threshold=0.62026.01 | 31.8 | 13.9 | 27.7 | 32.2 | 15.9 | 27.6 | |
| Sci-BERTThreshold Setting=T-2, Textual region threshold=0.7, Visual region threshold=0.62026.01 | 31.3 | 16.9 | 25.4 | 30.7 | 16.9 | 25.2 | |
| Sci-BERTThreshold Setting=T-3, Textual region threshold=0.6, Visual region threshold=0.62026.01 | 22.2 | 5.6 | 0.3 | 21.9 | 5.3 | 22 | |
| SPECTERThreshold Setting=T-2, Textual region threshold=0.7, Visual region threshold=0.62026.01 | 21.1 | 3.5 | 21.9 | 21.1 | 3.45 | 22 | |
| SPECTERThreshold Setting=T-3, Textual region threshold=0.6, Visual region threshold=0.62026.01 | 17.4 | 1 | 19.1 | 17.5 | 1.1 | 19.1 |