Scanned/Photographed Data Table
.png.jpg.jpeg.pdfAny camera, scanner, or PDF export of a printed table — lab notebooks, old papers, instrument printouts
What it is
A photo, scan, or PDF of a printed table of numbers — not a native data format at all. AltaiPlot runs OCR (text recognition) plus grid-structure detection to turn the image back into rows and columns you can plot.
Who produces it
Anyone digitizing an old lab notebook, a printed instrument readout, or a table from a paper/PDF that has no underlying digital data file.
What we read / what we don't read
What we read
- The table grid structure (rows/columns) detected from the image
- Cell text via OCR, parsed into numbers where possible
- The first page of a multi-page PDF
What we don't read
- OCR accuracy on real photographed tables is NOT perfect — benchmarked in the mid-50%s cell-accuracy range on a held-out real-photo test set, not the ~99% you'd get from a native file. Use the interactive review step (not the headless/automatic import) so you can correct misread cells before they land in a sheet
- Tables photographed at a steep skew angle (roughly beyond 8°) may misread a cell without any warning
- A table that spans more than one PDF page is not merged — only the first page is read
- Handwriting is not a target — this is built for printed/typed tables
How to import it into AltaiPlot
- Open AltaiPlot and drag your photo/scan/PDF onto the window, or use File → Import.
- AltaiPlot detects the image/PDF and opens the interactive scanned-table review panel — this is the recommended path, not the silent automatic one.
- Review the detected grid and OCR'd values, correcting any misread cells before committing.
- Click Import once the preview looks right.
Auditability note: this page's capability claims are sourced fromsrc/renderer/src/engine/v7/import/importers/ScannedTableImporter.ts:4 — see the full format list for every extension AltaiPlot recognizes.