How to adjust the formatting after the scanned document is converted to Word?

Key points of this article

After conversion, first check with the original to determine whether the problem is a column, paragraph or line break, and then decide on the revision method. This is more time-consuming than hard adjustments one by one.
When two-column text is strung into a long line, first use the section or column function to restore the number of columns, and then check whether the breakpoints between columns are correct.
Paragraphs are split into one line and one paragraph, usually caused by a hard carriage return in the original text. Unified processing using search and replace is faster than manual merging.
Typos and sentence fragmentation after conversion of the scanned document need to be checked manually. When revising the typesetting, important paragraphs should be read through to avoid handing over the errors as they are.

After converting the scanned document into Word, many people find that the text is still there, but the layout is messed up: the original double columns become one long series, paragraphs are broken line by line, words at the end of the line are split in half, and page numbers and headers are mixed into the text. This article is written for those who have obtained the conversion results and are currently modifying them. We will not discuss the settings before conversion. We will only deal with the generated Word files and tell you how to determine the problem, the order in which to repair it, and how to avoid rework after repairing.

After reading this article, you should be able to do three things: first, take a minute or two to distinguish whether the messed up columns, paragraphs or line breaks are; second, fix them in the order of easier first and then harder, without having to manually adjust each paragraph; third, know where you need to go back to the original to check. The same idea applies to documents converted from pictures (such as text photos taken with mobile phones).

The typesetting is messed up after conversion, how to fix it? One sentence conclusion

First, compare the originals to find out the confusing types, and then batch process them by type: fix column problems using section and column number settings, paragraph problems using search, replace and merge, and line break problems using a unified paragraph format. Manual adjustments section by section are usually time-consuming and can easily lead to chaos.

your situation	Recommended practices	reason
The original text was in two columns, but after conversion it became a long text in one column.	Divide sections by columns, reset the number of columns, and then check breakpoints	The column structure belongs to layout information, and modified paragraphs cannot be restored.
Each line is split into paragraphs, with large spaces between paragraphs	Use search and replace to remove redundant paragraph marks, and then unify the paragraphs before and after	The hard return in the original text is regarded as the end of the paragraph.
Words at the end of lines are split, or headers and page numbers are mixed into the text	Clean up abnormal line breaks and delete duplicate header and footer text	Treat layout elements as text content when identifying
The text is basically correct, with only partial misalignments.	Only select the misaligned parts and use the format brush to unify them into the main text format.	There is no need to redo the entire article for local problems
The scan itself is crooked or of poor quality	Do OCR in PDF first, then export to Word for modification	The quality of the text layer is poor. Directly repairing Word will make the repair more messy.

First determine what kind of problem it is

Before repairing, spend two minutes comparing the originals and classifying the messed up parts. The table below can help you locate it quickly.

symptoms	Common causes	How to confirm
The contents of the two columns are connected into one long article, and the text in the left and right columns are connected together.	The columns were not recognized during conversion and were sorted directly in reading order.	Count how many columns there are on each page compared to the original, and then see if there is only one column in Word.
Each line is followed by an end-of-paragraph mark and the sentence is truncated	The newline character in the original text is recognized as a hard return	Turn on Show edit marks to see if there are paragraph marks at the end of each line
There is too much space between paragraphs, or there are blank lines between the text.	The distance before and after the paragraph is not uniform, or there are multiple empty paragraphs in succession.	Select a paragraph to view the paragraph format and compare the spacing between adjacent paragraphs.
The words at the end of the line are broken, and the two lines before and after have half a word each.	Hyphen breaks in the original text are retained, or formatting and line breaks are treated as the main text.	Search for the same word in the original and Word to see if the disconnection position is consistent.
Headers, page numbers, and book titles appear repeatedly in the text	Page layout elements are recognized as body content	After turning to a few pages, observe whether the same text appears on every page.
The lines of the table disappear and the rows and columns are misaligned	The table is recognized as ordinary text, or the cells are broken up	Check how many rows and columns there are in the table in the original document, and compare them with those in Word

Check and repair item by item

It is recommended to deal with it in the following order, first fix the issues with a large impact, and then fix the local details. After completing one item in this way, the subsequent workload will be significantly reduced.

Keep an unmodified copy first. Save a copy of the conversion results before modification, so you can roll back if the changes are bad. This step may seem redundant, but it can save a lot of time in redoing it.
Confirm whether the text layer is available. Try to select a piece of text in Word or PDF. If you can select and copy it normally, it means that the text layer is available and you can directly modify the typesetting. If the selected text is a whole picture, or the text is garbled, you need to go back to the OCR process first.
Turn on Show Edit Marks to observe the paragraph structure. Turn on the option to show paragraph symbols in Word to see where the actual paragraph ends and where they are just line breaks caused by hard returns.
Unified cleaning of abnormal line breaks. Use find and replace, replace the hard returns with spaces, and re-mark the paragraph separations that need to be retained.
Restore the column structure. According to the number of columns on each page of the original text, use section breaks to separate the parts that need to be divided into columns, and then set the corresponding number of columns.
Delete headers, page numbers and repeated layout text. Delete redundant content in the header and footer areas, or use the search function to delete recurring header text in batches.
Unify the front and back paragraphs and line spacing. Select the main text, set it to the same paragraph format, delete consecutive empty paragraphs, and restore the rhythm of the layout to normal.
Read and check page by page. Finally, read it from the beginning, focusing on numbers, names, proper nouns, and disconnected words.

Column confusion: double columns become one long list

Improper column division is a problem that affects reading because the contents of the left and right columns are joined together, making it difficult for readers to see the boundaries of the paragraphs. When repairing, first find the starting and ending positions of the columns on each page, insert section breaks at the starting and ending positions, and then set the number of columns for the sections that need to be divided into columns. After the section breaks are inserted, the text in the columns will be distributed in order, but the breakpoints between columns are not necessarily consistent with the original text. You need to compare it page by page and manually move the sentences across the columns to the correct position. If the columns of the original text are inconsistent on different pages (such as a title column and a double column of text), they must be set in sections, and one number of columns cannot be used uniformly.

Paragraphs are broken up: one paragraph per line or paragraphs are glued

Paragraph problems usually appear in two ways. One is that the end of each line is the end of a paragraph, causing a sentence to be split into multiple paragraphs; the other is that the places that should be divided into paragraphs are connected together. The former can be handled with find and replace: replace the end of the paragraph with a space, and then manually re-break the real paragraphs. The latter requires checking paragraph by paragraph and adding paragraph end characters where they should be separated. There is an easily overlooked detail here: after the Chinese period or question mark, if it is followed by the content of the next paragraph, it is usually the position of the paragraph. Judging by the punctuation at the end of a sentence is more reliable than staring at it line by line.

Line breaks and word hyphenation: words at the end of lines are split, and headers are mixed into the text.

End-of-line word breaking is a typical problem after conversion of scanned documents. The reason is that the words that are broken at the end of the line in the original text are retained into two paragraphs of text. When repairing, first find the broken words in the original document, and then merge the broken parts in Word. If headers, page numbers and book titles appear in the main text, they need to be deleted and checked to see if there are any leftovers on each page. If there is a large amount of these contents, they can be processed collectively in the header and footer areas, which saves time rather than deleting them page by page. The converted text requires manual verification, especially numbers, names, and professional terms. Typos and sentence segmentation errors identified by OCR are often hidden in these places.

If your file is a scanned PDF, you can also perform OCR on the scanned document in the PDF editor of the OmniDoc desktop client, and then export it to Word. The resulting text layer is usually cleaner than a direct conversion and will have fewer typographical issues. After the modification is completed, you can use app.unidoc.top's UniDoc web editor to continue editing with Word compatibility, or use the PDF editing function in the download.omnidoc.top download desktop client.

How to avoid it happening again next time

The workload of repairing typesetting is often greater than the conversion itself, so making the following things a habit can significantly reduce rework.

After conversion, check three pages first and then revise them in full. By spot-checking the first page, middle pages, and last pages, you can quickly determine whether the confusion is an individual phenomenon or an overall problem, and avoid doing a lot of work in the wrong direction.
Record the steps for each modification. Write down the effective search and replacement conditions and section locations in your notes so that you can apply them directly next time you encounter similar files.
Save one version before and after modification. The version number is clear, so there will be no confusion when rolling back, and it is also easy to compare with colleagues.
Treat checking as a finishing move. Fixing the typesetting does not mean that the content is correct. Numbers, proper nouns and broken words must be compared with the original.

❓ Frequently Asked Questions (FAQ)

1. The format of the scanned document is messed up after being converted to Word. Can it be restored with one click?

Usually not. The layout information of the conversion result has been lost or misjudged, and the one-click function can only solve part of the problem. A more practical approach is to first determine the type of confusion, and then use sectioning, search and replacement, and paragraph formatting to repair them uniformly.

2. How to adjust the format of the image after it is converted into a document?

The idea is the same as that of scanned documents: first check whether the text can be selected, and then determine whether it is a problem of columns, paragraphs or line breaks. If the image itself is skewed or blurred and the repair space is limited, you can perform OCR on the image first and then modify the text layer.

3. How to keep two columns after converting a two-column document to Word?

In Word, according to the number of columns on each page of the original text, use section breaks to separate the parts that need to be divided into columns, and then set the number of columns for each section. The title or banner part must be a separate section and cannot share the same number of columns with the double-column text.

4. The converted text cannot be selected, what’s going on?

This usually means that the file is a scanned image, with no text layer available. You can do OCR in PDF first, generate a text layer and then export it to Word, and pay attention to check the recognition results.

5. How to quickly find typos when editing typesetting?

After the typesetting is completed, focus on reading through the numbers, names, units and professional terms, and compare them with the originals one by one. OCR easily confuses similar words. The probability of errors in these places is higher, and the benefits of verification are also high.

📝 Summary

If the layout of the scanned document is messed up after being converted to Word, first compare it with the original document to determine whether it is a column, paragraph or line break problem, and then fix it in the order of easier first and then more difficult. This can usually reduce rework to a minimum. After completing the revision, don’t forget to read through it, especially numbers and proper nouns, because the OCR results require manual confirmation.

Open OmniDoc, edit and export documents online