Convert JPG to CSV

Excel, database, and analysis software require rows and fields rather than a photograph of a table, receipt, form, or list. Converting the image makes that information editable and searchable; changing .jpg to .csv does not create structured data.

What the JPG format is

JPG, also called JPEG, is a raster image format used by cameras, scanners, smartphones, screenshots, and image-editing software. Its lossy compression keeps photographic and scanned files relatively small, but the pixels do not inherently contain text, fields, rows, or columns. A JPG can display a table without storing the table as data.

What the CSV format is

CSV is a plain-text representation of tabular data. Lines normally represent records, while commas, semicolons, or another delimiter separate fields; fields containing delimiters, quotes, or line breaks must be quoted. Spreadsheet applications, databases, statistical programs, and scripts can import it.

CSV is small, portable, and easy to process, but it does not preserve fonts, page layout, images, formulas, multiple worksheets, or dependable cell types. Use UTF-8 when non-ASCII characters must survive transfer, and confirm the delimiter expected by the receiving program.

How to convert a table in a JPG

Use Excel for a clear printed table

1. Open a Microsoft 365 version of Excel that includes the image-recognition feature.
2. Choose Data → From Picture → Picture From File, then select the JPG. On supported mobile versions, use the equivalent image-to-data command.
3. Review Excel's detected cells before inserting them, correcting column boundaries, merged cells, dates, decimal separators, negative values, blank cells, and leading zeros.
4. Save the corrected workbook first if you need an editable audit copy.
5. Choose File → Save As and select CSV UTF-8 (Comma delimited) (*.csv). A CSV export normally contains only the active worksheet.

Use ABBYY FineReader PDF for difficult scans

1. Open the JPG in ABBYY FineReader PDF.
2. Run OCR with the document's correct language and enable table recognition.
3. Inspect the recognized table, especially merged cells, column boundaries, reading order, and characters confused by the scan.
4. Export to XLSX, open the result in a spreadsheet application, and use File → Save As to create a UTF-8 CSV. Direct CSV export depends on the installed FineReader edition.

Process batches with OCR software or an API

Tesseract OCR can extract text but does not reliably reconstruct complex tables by itself. For example, tesseract input.jpg output -l eng creates output.txt; a separate parser must identify rows and columns and write valid CSV. For repeated table extraction, AWS Textract, Google Cloud Document AI, and Azure AI Document Intelligence provide table-aware OCR; a script can convert their returned cells to CSV.

For occasional, non-sensitive images, browser services such as OCR.space and OnlineOCR can recognize text or export spreadsheet-compatible results. Open the recognized result in a spreadsheet program, correct it, and save it as CSV. Do not upload identity documents, financial records, confidential receipts, or proprietary material to an online service.

Convert pixels instead of text

If the intended result is a numeric matrix containing every JPG pixel, OCR is the wrong method. Read the image with Python, MATLAB, or another image-processing tool and write pixel values to CSV; a color image normally produces separate red, green, and blue values, and the resulting file can be much larger than the JPG. For example, Python can use Pillow and NumPy to load the image and numpy.savetxt to write a grayscale array, but CSV is usually less efficient than a binary array format for this purpose.

Quality and compatibility limits

OCR accuracy depends on resolution, contrast, skew, blur, handwriting, font, compression artifacts, and table lines. Use a straight, tightly cropped image; for a scanned document, about 300 ppi is a useful target. Correct rotation and improve contrast before recognition when necessary.

Compare every output row with the source image. OCR can change 0 to O, 1 to I, alter decimal points and currency symbols, and remove leading zeros. Regional spreadsheet settings may interpret dates differently or expect semicolons instead of commas.

Retain the original JPG and corrected XLSX when traceability matters. CSV export discards formatting, formulas, additional worksheets, and much of the source application's type information.