See the difference on real scans

Nine scans, converted to Word by a leading desktop OCR program and by UnfoldPDF. Nothing edited.

1 of 9

Budget figures you can edit

Expense form with budget boxes and a ruled table

The scan What both programs got
The scanned page: Expense form with budget boxes and a ruled table
A leading desktop OCR program 16 frames, 1 text box
The desktop OCR result, as shown in WordThe desktop OCR result, as shown in Word with formatting marks
UnfoldPDF No frames or text boxes
The UnfoldPDF result, as shown in WordThe UnfoldPDF result, as shown in Word with formatting marks

Select a page to open it full size.

UnfoldPDF did better

  • Amounts are editable text ($7,569,000.00). The desktop program kept them as pictures.
  • “Current Budget”, not “Current Rudgcl 1”.
  • Charge codes sit in a real Word table.

UnfoldPDF still gets wrong

  • Signatures become stray text (“123/6”), and one “Date” reads “Dame”.

2 of 9

A typed form in two columns

1976 typed form in two columns

The scan What both programs got
The scanned page: 1976 typed form in two columns
A leading desktop OCR program 155 frames, 2 text boxes
The desktop OCR result, as shown in WordThe desktop OCR result, as shown in Word with formatting marks
UnfoldPDF No frames or text boxes
The UnfoldPDF result, as shown in WordThe UnfoldPDF result, as shown in Word with formatting marks

Select a page to open it full size.

UnfoldPDF did better

  • “20,000 cigts.”, not “36~7OOO ciqts.”
  • “Cigarettes”, not “Ciqarettos”.
  • Each label sits next to its value in a table.

UnfoldPDF still gets wrong

  • “lbs.” reads “1bs.”, and “Shipping”, under a hole punch, reads “epping”.

3 of 9

Score tables with significance marks

1982 test memo with score tables

The scan What both programs got
The scanned page: 1982 test memo with score tables
A leading desktop OCR program 19 frames, 4 text boxes
The desktop OCR result, as shown in WordThe desktop OCR result, as shown in Word with formatting marks
UnfoldPDF No frames or text boxes
The UnfoldPDF result, as shown in WordThe UnfoldPDF result, as shown in Word with formatting marks

Select a page to open it full size.

UnfoldPDF did better

  • Labels sit beside their values, not in separate text boxes.
  • Scores keep their asterisks (“3.60 **”). The desktop program has “3e64***”.
  • “MENTHOL PRODUCT TEST”, not “MENTHOL 匸R0DXT TEST”.

UnfoldPDF still gets wrong

  • “PAGE # 1 of 4” loses its “1”.
  • Scale notes join the next row’s label, and one row lands in the wrong columns.

4 of 9

Two wide tables and a tick box

Report with tick boxes and two six-column tables

The scan What both programs got
The scanned page: Report with tick boxes and two six-column tables
A leading desktop OCR program 70 frames, 2 text boxes
The desktop OCR result, as shown in WordThe desktop OCR result, as shown in Word with formatting marks
UnfoldPDF No frames or text boxes
The UnfoldPDF result, Word page 1 of 2Page 1The UnfoldPDF result, Word page 1 of 2 with formatting marksPage 1
Word page 2 of 2The UnfoldPDF result, Word page 2 of 2Page 2The UnfoldPDF result, Word page 2 of 2 with formatting marksPage 2

Select a page to open it full size.

UnfoldPDF did better

  • Both six-column tables are real Word tables.
  • The ticked box comes through: “JUNE 30 ☑”.
  • “(15 + STORES)”, not “(45 + STORES)”.

UnfoldPDF still gets wrong

  • “Retail” reads “Retall”.
  • Empty rows push the last row onto a second page.

5 of 9

A handwritten note and a hidden number

Form with a handwritten note and a blacked-out number

The scan What both programs got
The scanned page: Form with a handwritten note and a blacked-out number
A leading desktop OCR program 42 frames
The desktop OCR result, as shown in WordThe desktop OCR result, as shown in Word with formatting marks
UnfoldPDF No frames or text boxes
The UnfoldPDF result, as shown in WordThe UnfoldPDF result, as shown in Word with formatting marks

Select a page to open it full size.

UnfoldPDF did better

  • The handwritten note becomes text. The desktop program has “)uZL-”.
  • The blacked-out number shows as “[Redacted]”.
  • “☐ Yes ☐ No ☐ Pending”, not “Qy®s ONo QPending”.

UnfoldPDF still gets wrong

  • The handwritten “Note:” reads “hate:”.

6 of 9

A form filled in by hand

Job application filled in by hand (sample with made-up details)

The scan What both programs got
The scanned page: Job application filled in by hand (sample with made-up details)
A leading desktop OCR program 65 frames
The desktop OCR result, as shown in WordThe desktop OCR result, as shown in Word with formatting marks
UnfoldPDF No frames or text boxes
The UnfoldPDF result, Word page 1 of 2Page 1The UnfoldPDF result, Word page 1 of 2 with formatting marksPage 1
Word page 2 of 2The UnfoldPDF result, Word page 2 of 2Page 2The UnfoldPDF result, Word page 2 of 2 with formatting marksPage 2

Select a page to open it full size.

UnfoldPDF did better

  • Handwriting becomes text: “ROMEO GODFRIED WALTER”, not “KOmEO Lionel ED”.
  • Crosses in boxes become ticked boxes (☑).
  • Each phone number stays on one line with its label.

UnfoldPDF still gets wrong

  • “NAMIBIAN” comes out in Cyrillic, and one “WINDHOEK” as “KINDHOEK”.
  • A stray ☑ in the instructions; the form runs onto a second page.

7 of 9

Tick boxes and dates beside each entry

1982 summary with tick boxes and dated entries

The scan What both programs got
The scanned page: 1982 summary with tick boxes and dated entries
A leading desktop OCR program 15 frames, 4 text boxes
The desktop OCR result, as shown in WordThe desktop OCR result, as shown in Word with formatting marks
UnfoldPDF No frames or text boxes
The UnfoldPDF result, as shown in WordThe UnfoldPDF result, as shown in Word with formatting marks

Select a page to open it full size.

UnfoldPDF did better

  • The ticked box survives: “COMBINED ☑”. The desktop program lost it.
  • Dates stay on their entries’ rows, not in text boxes.
  • “COMPOUND”, not “COMPOUNO”.

UnfoldPDF still gets wrong

  • Signatures become text (“Jimmy J. Bull”).

8 of 9

A boxed lab form

1983 boxed lab form with tick boxes

The scan What both programs got
The scanned page: 1983 boxed lab form with tick boxes
A leading desktop OCR program 2 frames
The desktop OCR result, as shown in WordThe desktop OCR result, as shown in Word with formatting marks
UnfoldPDF No frames or text boxes
The UnfoldPDF result, as shown in WordThe UnfoldPDF result, as shown in Word with formatting marks

Select a page to open it full size.

UnfoldPDF did better

  • “☑ REFRIGERATOR 8℃”, not “OREFRI6ERAT0R”.
  • “VIRGINIA”, not “VlPGlNtA”.
  • “is attached”, not “1s attached”.

UnfoldPDF still gets wrong

  • “PROJECT NO.” comes after its number.
  • The signature becomes text (“D. Stato”).

How these were made

Try it on your own page

Your first page is free, no account needed.

Choose a file