YogeLiu/doc_processer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: update threshold

Browse Source
This commit is contained in:
liuyuanchuang
2026-02-07 13:26:57 +08:00
parent de66ae24af
commit d86107976a
2 changed files with 1 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
app/services/layout_detector.py
View File

@@ -134,7 +134,7 @@ class LayoutDetector:
) )
) )
mixed_recognition = any(region.type == "text" and region.score > 0.6 for region in regions) mixed_recognition = any(region.type == "text" and region.score > 0.3 for region in regions)
return LayoutInfo(regions=regions, MixedRecognition=mixed_recognition) return LayoutInfo(regions=regions, MixedRecognition=mixed_recognition)

5
app/services/ocr_service.py
View File

@@ -701,11 +701,6 @@ class MineruOCRService(OCRServiceBase):
if "results" in result and "image" in result["results"]: if "results" in result and "image" in result["results"]:
markdown_content = result["results"]["image"].get("md_content", "") markdown_content = result["results"]["image"].get("md_content", "")
# Check if markdown contains formula image references
if "![](images/" in markdown_content:
# Use PaddleOCR-VL to recognize the formula
markdown_content = self._extract_and_recognize_formulas(markdown_content, image)
# Apply postprocessing to fix OCR errors # Apply postprocessing to fix OCR errors
markdown_content = _postprocess_markdown(markdown_content) markdown_content = _postprocess_markdown(markdown_content)