Extract a table from JPG
Use a high-quality JPG when the table came from a camera or scanner. Heavy compression can blur small numbers and punctuation.
Extract Table From an Image Online
Upload or paste a JPG, PNG, WEBP, BMP, or screenshot. Picture2Txt proposes editable rows and columns, then lets you download the reviewed table as CSV or Excel.
Drop a clear table image here
Use a straight image with readable headers and consistent row or column alignment.
Up to 20 MB•Clear headers work best•Review every value
Editable reconstructionCorrect rows, columns, and cell text
Spreadsheet exportDownload CSV or XLSX after review
Position-aware readingUses recognized words and bounding boxes
Table extraction workflow
How to convert an image to a table online
Choose a readable source
Upload a table screenshot, a high-quality JPG, or a sharp PNG. Keep the camera parallel to printed pages.
Select the table language
Choose the language model that matches the headers and cell text.
Run the table extractor
The browser recognizes words, groups lines into candidate rows, and estimates repeated column positions.
Correct the editable grid
Fix merged columns, split values, add missing rows, and verify totals before export. If the first layout is wrong, set the expected column count or adjust cell-gap sensitivity and rebuild the grid without reading the image again.
Source preparation
Improve results before table extraction
- Crop unrelated captions, logos, and paragraphs around the table.
- Use a straight image without strong perspective distortion.
- Keep headers and body text large enough to show complete strokes.
- Reduce glare and shadows across printed rows.
- Use the image text enhancer for faint contrast.
- Use the screenshot to text tool when you need plain text rather than a grid.
Different source formats
Extract tables from JPG, PNG, and screenshots
The best input is the clearest one. File format matters less than readable text, straight rows, and stable spacing.
Extract a table from PNG
PNG often keeps screenshot text and grid edges sharp. Transparent backgrounds can still reduce contrast if the viewer background changes.
Copy a table from a screenshot
Crop the screenshot to the grid. Remove toolbars and side panels so the table extraction engine sees fewer competing text regions.
How it works
What the table extractor reads
The browser uses Tesseract.js to recognize words and their bounding boxes. Picture2Txt clusters nearby words into lines and uses recurring horizontal positions to propose columns.
This method works best with simple printed tables. It is not a full document-layout engine, so the editable grid is an informed draft rather than an exact copy.
Export choices
Use the image to CSV or image to Excel output
CSV stores plain cell values in a widely supported text format. XLSX creates a spreadsheet workbook from the corrected grid. Both exports happen after you review the table.
The browser builds the files locally. MDN explains how browser applications can read user-selected files through the File API.
Table extraction does not preserve every layout detail. Merged cells, formulas, borders, colors, nested tables, multi-line headers, and handwritten values may be incorrect. Treat the output as an editable reconstruction and verify every important value.
Useful details
Common questions about image table extraction
These short explanations cover common tasks, limits, and result-checking steps.
Can I extract a table from an image online?
Yes. Upload a clear table image, start recognition, correct the proposed grid, and export the reviewed values.
Can a JPG become an editable table?
Yes. The table extractor can recognize text in a JPG and propose rows and columns for manual correction.
Can I extract a table from PNG files?
Yes. PNG images often preserve sharp text and grid lines, which can help recognition when the source is clear.
Can I convert an image to CSV?
Yes. After reviewing the grid, download the table as a CSV file directly in the browser.
Can I convert an image to Excel?
Yes. The tool creates an XLSX workbook from the corrected cell values without a server export step.
Does table extraction preserve merged cells?
Not reliably. Merged cells, nested headers, and complex spans usually need manual correction.
Can the image to table converter read borders?
It uses recognized word positions more than visual borders. Clear alignment still improves row and column detection.
Can I copy a table from an image?
Yes. Correct the cells, then copy tab-separated table data for a spreadsheet or document.
Does a borderless table work?
The tool can infer columns from repeated text positions, but uneven spacing may merge or split cells.
Can I edit every extracted cell?
Yes. Each cell remains editable before CSV or XLSX export.
Can I correct the detected columns without reading the image again?
Yes. Choose a fixed column count, adjust cell-gap sensitivity, and rebuild the current grid without rerunning text recognition.
Is the table image uploaded?
Text recognition and spreadsheet export are designed to run inside the active browser workflow.
Should totals and amounts be checked?
Yes. Verify totals, percentages, dates, decimal points, and currency values against the source image.
Frequently asked questions
More answers about this tool
How does the image to table converter find columns?
It recognizes words with bounding boxes, groups words into lines, and compares repeated horizontal positions to propose column boundaries.
Why did two table columns merge?
Small gaps, perspective distortion, uneven alignment, faint text, or long cell content can make separate columns appear connected.
Does XLSX export preserve the original table design?
No. The workbook contains the corrected cell values. It does not reproduce exact fonts, colors, borders, formulas, or merged cells.
What image gives the best table extraction result?
Use a straight, sharp image taken parallel to the table. Keep headers readable and avoid shadows, glare, and heavy compression.
Can I change the proposed column structure?
Yes. Select an expected column count, adjust the cell-gap sensitivity, and rebuild the grid. Lower sensitivity splits cells more often, while higher sensitivity keeps nearby words together.
Can I import the result into Google Sheets?
Yes. Download CSV or XLSX, then import the file into Google Sheets. Review the cell structure before using formulas or totals.