File tree Expand file tree Collapse file tree 2 files changed +4
-4
lines changed
Expand file tree Collapse file tree 2 files changed +4
-4
lines changed Original file line number Diff line number Diff line change @@ -145,7 +145,7 @@ Validated on 10 manually annotated scientific posters:
145145| Word Capture | 0.96 | ≥0.75 |
146146| ROUGE-L | 0.89 | ≥0.75 |
147147| Number Capture | 0.93 | ≥0.75 |
148- | Field Proportion | 0.99 | 0.30 –2.50 |
148+ | Field Proportion | 0.99 | 0.50 –2.00 |
149149
150150** Pass Rate** : 10/10 (100%)
151151
Original file line number Diff line number Diff line change @@ -11,15 +11,15 @@ The pipeline is validated using four complementary metrics:
1111| ** Word Capture (w)** | Proportion of reference vocabulary in extracted text | ≥0.75 | Measures lexical completeness |
1212| ** ROUGE-L (r)** | Longest common subsequence similarity | ≥0.75 | Captures sequential text preservation |
1313| ** Number Capture (n)** | Proportion of numeric values preserved | ≥0.75 | Validates quantitative data integrity |
14- | ** Field Proportion (f)** | Ratio of extracted to reference JSON elements | 0.30 –2.50 | Accommodates layout variability |
14+ | ** Field Proportion (f)** | Ratio of extracted to reference JSON elements | 0.50 –2.00 | Accommodates layout variability |
1515
1616### Pass Criteria
1717
1818A poster passes validation if ALL conditions are met:
1919- Word Capture ≥ 0.75
2020- ROUGE-L ≥ 0.75
2121- Number Capture ≥ 0.75
22- - Field Proportion between 0.30 and 2.50
22+ - Field Proportion between 0.50 and 2.00
2323
2424## Metric Implementation
2525
@@ -72,7 +72,7 @@ reference_fields = count_json_fields(reference_json)
7272field_proportion = extracted_fields / reference_fields
7373```
7474
75- The extended range (0.30 –2.50 ) accommodates:
75+ The extended range (0.50 –2.00 ) accommodates:
7676- Nested vs flat section structures
7777- Variable poster layouts
7878- Optional metadata fields
You can’t perform that action at this time.
0 commit comments