Capture three sources, merge OCR output, validate values 1–16.
Designed for split datasets where the first image holds the main table and the second and third images contribute missing details. Manual timestamps and operator notes stay attached to the final combined record.
Captured images
0
Values found
0/16
Confidence
0%
Capture manager
Image inputs
Upload up to three images. Recommended order: main dataset, detail extraction, additional detail extraction.
Manual input module
Timestamps & metadata
Confidence engine
Extraction confidence
0%
Low confidence
Quality from OCR + completeness
Once analysis finishes, this panel shows confidence scoring for OCR quality, completeness, conflict management, and sequence consistency.
Pipeline control
Run multi-image OCR
Image1 → OCR → Image2 → OCR → Image3 → Aggregate → Validate → Score.
No capture set loaded
Start with your own images or load the built-in demo set. The demo includes a realistic archive-style table split across three images containing values 1 through 16.