How to Digitize Paper Tables into Excel Without Retyping a Single Cell

Somewhere in your office, classroom, or kitchen drawer, there is a printed table you wish you had as a spreadsheet. A supplier price list. A professor's grading rubric photocopied from a textbook. An old inventory count scribbled on lined paper. Retyping it all feels like punishment. OCR tools now handle this in seconds.

Here is how to go from paper table to editable XLSX file, with practical tips that determine whether you get clean data or a mess you still have to fix by hand.

Step 1: Get a Good Photo

The single biggest factor in extraction accuracy is the image itself. A blurry, shadowed phone photo of a crumpled page will produce garbled results no matter which tool you use.

What works well:

  • Flat surface, even lighting, no shadows crossing the table
  • Camera held directly overhead (not at an angle)
  • The entire table visible in frame, with a small margin on each side
  • Screenshots of on-screen tables (these convert almost perfectly because the text is already rendered cleanly)

What causes problems:

  • Handwriting (especially cursive or smudged ink)
  • Glossy paper with reflections
  • Tables where gridlines are missing and columns rely on whitespace alignment alone

If you are digitizing from a physical book, press the spine flat and photograph one page at a time. Cropping out everything except the table before uploading also helps the OCR model focus on the right region.

Step 2: Run It Through an Image-to-Excel Converter

Upload your JPG, PNG, GIF, or WebP file to Imgtosheet. The tool sends your image through an OCR model that reads the text and reconstructs the original rows and columns. You get a preview of the extracted table in your browser, then download it as an XLSX file.

A few things worth knowing about the process:

  • Files can be up to 10 MB each.
  • The XLSX file is generated in your browser. Your image is not stored on any server or database after processing.
  • The free tier handles one image at a time. If you are working through a stack of documents, you will upload them individually (a batch mode for 5 to 20 images in a single workbook is planned for a future Pro tier).
  • The output is an XLSX file, which opens in Excel, Google Sheets, or Apple Numbers. If you need CSV, open the XLSX and export from there.

Step 3: Check the Output Before You Trust It

OCR is good, but it is not infallible. Always scan the preview before downloading. Common things to watch for:

  • Merged cells that got split into separate columns
  • Numbers read as text (a leading apostrophe or invisible formatting can prevent formulas from working)
  • Currency symbols or special characters that got dropped or substituted
  • Row alignment shifts where one row's data slid into the wrong columns

Fixing two or three cells is still faster than retyping 200 of them. But catching errors before the file enters your workflow saves you from building calculations on bad data.

Step 4: Organize the Workflow When You Have Many Documents

Digitizing one table is straightforward. Digitizing 30 over the course of a week, while juggling other work, is where things fall apart. You lose track of which pages you have already converted, which files still need cleanup, and which ones are ready to merge into a master sheet.

A simple kanban board solves this. Create three columns: "To Scan," "Needs Review," and "Done." Each card represents one document. Move cards across as you work through the stack, whether that is a sticky-note board, a spreadsheet, or any lightweight task tool you already use.

This structure also helps if multiple people are splitting the work, since everyone can see at a glance which documents still need attention.

When OCR Beats Manual Entry (and When It Does Not)

OCR shines on printed text with clear gridlines: invoices, price sheets, report tables, textbook data. For these, extraction is fast and accuracy is high.

It struggles with:

  • Handwritten tables (legibility varies too much)
  • Tables embedded in complex page layouts with sidebars, footnotes, and overlapping elements
  • Very small text in high-density tables (zooming in and cropping to sections helps)

For a 50-row printed table with standard fonts, you will save 15 to 30 minutes of manual typing. For a 10-row handwritten table with neat block letters, OCR might still save you time, but expect to correct a few cells. For a 10-row table with messy cursive, you might be faster just typing it.

The Short Version

Photograph the table flat and well-lit. Upload the image. Check the preview for errors. Download the XLSX. If you are processing many documents, track them on a board so nothing gets lost. That is the whole workflow, and it replaces hours of tedious data entry with a few minutes of setup and review.