Skip to content

OCR parity: markitdownnet tuned (eng, PSM=6, DPI=300) #25

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
mapo80 wants to merge 4 commits into
base: main
Choose a base branch
from
Open

OCR parity: markitdownnet tuned (eng, PSM=6, DPI=300) #25

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
cdeb844
OCR parity: markitdownnet tuned (eng, PSM=6, DPI=300)
mapo80 Aug 21, 2025
238f9e0
OCR parity: remove pre-binarization, gated deskew, DPI metadata; rege…
mapo80 Aug 21, 2025
e591c57
OCR parity: remove pre-binarization, gated deskew, DPI metadata; rege…
mapo80 Aug 21, 2025
990555b
OCR parity: disable pre-binarization, gate deskew (≥2°), set DPI meta…
mapo80 Aug 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 9 additions & 2 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,7 +64,7 @@ var options = new MarkItDownOptions
{
OcrDataPath = "/usr/share/tesseract-ocr/5/tessdata",
OcrLanguages = "eng",
PdfRasterDpi = 300
OcrUserDpi = 300
};
var converter = new MarkItDownConverter(options);
var result = await converter.ConvertAsync("sample.pdf", "application/pdf");
Expand All @@ -77,7 +77,7 @@ Console.WriteLine(result.Markdown);

* `OcrDataPath` – location of Tesseract language data (`TESSDATA_PREFIX`)
* `OcrLanguages` – languages passed to Tesseract (e.g. `ita+eng`)
* `PdfRasterDpi` – DPI for rasterising PDFs during OCR fallback
* `OcrUserDpi` – DPI for rasterising PDFs during OCR fallback
* `MinimumNativeWordThreshold` – minimum words before OCR is triggered
* `NormalizeMarkdown` – toggle Markdig normalisation

Expand Down Expand Up @@ -108,6 +108,13 @@ Docling's image samples are distributed as TIFF files. The comparison tool conve

These large arXiv PDFs showed a 99.37% word match rate and a 10.74% mean absolute error in bounding boxes.

## OCR benchmark

```bash
dotnet run --project tools/OcrBench -- extract --input-dir dataset/validation --out-dir dataset/validation/_ocr --threads 1 --langs eng --psm 6 --refresh markitdownnet
dotnet run --project tools/OcrBench -- compare --ocr-dir dataset/validation/_ocr --out-json artifacts/validation/OCR/bench-ocr.json --out-md artifacts/validation/OCR/summary-ocr.md
```

## Docling comparison

The `tests` project verifies Markdown and bounding box accuracy against the [Docling](https://github.com/docling-project/docling) ground truth for `ocr_test.pdf`.
Expand Down
Loading