OCR.hilllily.com | Multilingual OCR | Multy Column Support

1 Document Input

No file loaded

Click to upload or drag & drop

Multi-Page PDF, JPG, PNG, WEBP, TIFF, BMP (Dictionary Scans, Multi-Column Books)

Page Range

Render Resolution

PDF Engine

2 Multi-Language Script Matrix

eng

One-Click Mixed-Document Presets:

Primary Script / Language

Simultaneous Secondary (+)

Tip: For bilingual dictionaries like Molesworth's Marathi & English Dictionary, keeping mar+eng loads both Devanagari and Latin LSTM neural networks simultaneously so every line can switch scripts seamlessly.

3 Multi-Column Layout Engine

Detect Spanning Top Header First Extracts full-width title/running header before Col 1→2→3

Erase Vertical Divider Lines Prevents vertical column ink lines from becoming '|', '1', or 'I'

Column Gutter Sensitivity Medium-High (65%)

4 Vision Pre-Processor & Speed

Image Enhancement Filter

Parallel OCR Workers

Join End-of-Line Hyphens Clean Margin/Rule Noise

Upload a PDF or Image (or load the 3-Col Sample) to begin. 0%

Detected Cols -

Mean Confidence -

Words Extracted 0

Elapsed Time 0.0s

Interactive Multi-Column Layout & Projection Inspector

Awaiting Document

Page 1 / 1

Horizontal Ink Density Profile & Detected Column Gutters (Drag orange markers to fine-tune column boundaries): Header Split: Auto

Multi-Column Visual Segmentation Canvas

Upload your multi-column image or PDF. The engine will automatically detect any spanning header and split Column 1 → Column 2 → Column 3 for strict sequential reading order.

0. Spanning Header (Extracted 1st) 1. Column 1 (Extracted 2nd) 2. Column 2 (Extracted 3rd) 3. Column 3+ (Extracted Sequentially)

Drag vertical orange split lines or horizontal purple header line to adjust

Include [Column N] Markers