Is it okay to take a photo of the scanned copy? Photo quality self-check list

Key points of this article

Taking a picture with a mobile phone is not the same as a scan, but as long as the picture is clear, the paper is flat, and the four sides are intact, the photo taken can be used for OCR recognition.
Before taking pictures, check whether the light is even, whether the paper is flat, and whether the page is completely integrated into the picture. These three items determine most of the recognition quality.

-Multiple page files are first merged into one PDF and then recognized, which is more convenient for unified checking than processing them one by one.

OCR results may not be completely accurate. Numbers, names, professional terms and table contents must be checked against the original one by one.

Many people have a paper contract, a printed paper or an invoice in their hands. They want to copy the text on it and change it. They subconsciously take a picture with their mobile phone. After taking the photo, I discovered that the copied content was missing words, lines were broken, or simply couldn’t be selected. The problem is often not with the identification tool, but with the photo itself.

This article is for people who want to take pictures with their mobile phones and then transfer the text in the pictures. After reading this, you will be able to judge whether the photo you have is a "sufficient scan", know where to check before taking a photo, and master the complete steps of using OmniDoc to OCR a photo.

Is it okay to take a photo of the scanned copy? One sentence conclusion

Taking a photo itself is not equivalent to a scan, but as long as the photo meets the three conditions of being clear, flat, and complete, it can be treated as a scan and submitted to OCR for recognition. Successful recognition depends on the conditions when taking the photo, not the phone model. The table below gives methods for common situations.

your situation	Recommended practices	reason
Single page printing document with sufficient lighting	Lay it flat to take photos, convert directly to PDF and do OCR	Clear text and few recognition errors
Multi-page contract or paper	Take photos page by page, merge them into one PDF and then recognize them	Unified processing to facilitate overall verification of page numbers and order
The paper is creased or slightly curled	Flatten it with a book or tablet and take the photo again to avoid reflections.	Bending will deform the line spacing, and OCR will easily miss characters.
I couldn't take the picture clearly, the font is very small	Shoot closely but make sure all four sides are in the frame, or upload a scanned copy directly	When the resolution is insufficient, no matter how good the recognition is, the details cannot be restored.
text on screen	Prioritize taking screenshots instead of taking screen shots	Taking screen shots will produce moiré patterns, making screenshots cleaner.

Before you start: Confirm these points first

Before you start taking pictures, take a minute to check the following items. They do not require professional equipment, but each item directly affects the recognition results.

Whether the light is uniform. Shoot next to a window or under a desk lamp, so that the light shines evenly from the side or above. If the shadow of the phone falls on the paper, the edges of the text will be blurred, and OCR will misjudge the shadow as a stroke.
Whether the paper is flat. Curved edges, dog-ears, and bindings can deform the text near the spine. Flatten the book, or put a hard board under the paper, the effect is obvious.
Whether the page is completely pictured. All four edges should be included, and no headers, footers, or page numbers should be cut off. Text with cut-off edges is often missing after recognition and is not easily noticeable.
Whether the screen is facing the paper. Keep your phone as parallel to the paper as possible and avoid taking pictures at an angle. Shooting at an angle will make the text appear larger near and smaller at a distance, and the spacing between lines will also be distorted.
Whether there are reflections and finger occlusion. The highlight areas of glossy paper will eat up the writing, so you have to move your fingers and the shadow of your phone when taking photos.

Steps to perform OCR on photos taken in OmniDoc

The following steps are to organize the mobile phone photos into PDF, and then use the OmniDoc desktop client for OCR recognition.

Take photos page by page according to the self-check list above, one photo for each page, and make sure all four sides are complete and the text is clear before continuing.
Transfer the photos to your computer in page order and merge them into a PDF file using a common method. When there are many pages, first check whether the order is correct.
Open download.omnidoc.top to download the OmniDoc desktop client and follow the prompts to install it on your Windows, macOS or Linux computer.
Start the client, open the merged file with a PDF editor, and confirm that each page can be displayed normally.
Select the OCR function in the PDF editor to perform text recognition on the scanned pages.
After the recognition is completed, try to select a piece of text and copy it to confirm that it has become selectable text and not a picture.
When you need to modify it, edit the paragraph text directly in the PDF editor; when you need to continue using it for office work, you can export it to Word or Excel.
If the original document is a foreign language material that needs to be translated, you can also upload a scanned PDF on the OmniDoc web page, select bilingual comparison or translation only, and download the translation results.

Check each item after completion

OCR is not a one-click, absolutely correct step. After getting the results, it is recommended to check them one by one according to the table below.

Check items	How to see	What to do if there is a problem
Number of pages and order	Compare the originals to confirm that the content on each page is present and the order has not changed.	Go back to the merge step, reorder, and do OCR again
Numbers and amounts	Compare amounts, dates, and numbers one by one, especially easily confused characters such as 0 and 8, 1 and 7	Manual corrections are made against the originals, and important figures need to be checked twice.
Professional terms and names	Check rare words, English abbreviations and proper nouns	Modify in the editor and unify the full text writing
tables and paragraphs	Check whether the columns and columns are aligned and whether the paragraphs are mistakenly broken	Manually merge or split paragraphs using editing features
Missing words and missing lines	Check for missing page edges and footers	If there are missing words at the edges, it means the content was cut off when taking the photo, and the page needs to be retaken.

Machine recognition may make mistakes. When important content such as contract terms, medical records, financial data, etc. is involved, be sure to check it verbatim against the original.

🎯 Which method is suitable for you?

Only want to copy a paragraph: After taking a picture of this page, do an OCR and copy the required paragraph without processing the entire page.
To change it into an editable document: Merge the entire document into PDF, OCR and export it to Word, and then modify it in office software.
To understand foreign language materials: The scanned PDF can be directly uploaded to the OmniDoc web page for translation. Select bilingual comparison to facilitate comparison of the original text paragraph by paragraph.

❓ Frequently Asked Questions (FAQ)

1. How to convert pictures to text on mobile phone?

Take photos according to the list on your mobile phone, transfer the photos to your computer, merge them into PDF, and then use the OCR function of the OmniDoc desktop client to recognize them. If there are only one or two pictures with not much content, you can take a clear picture on your mobile phone first and then transfer it to the computer for processing. The operation will be more stable.

2. Can mobile phone screenshots be directly OCRed?

Screenshots are usually clearer than photos and are suitable for text on the screen, such as web pages, chat logs or parts of electronic files. After taking screenshots, you can also organize them into PDF and then recognize them. Just pay attention to whether the screenshots are complete and whether the font size is large enough.

3. There are many errors in the words recognized after taking the photo. What is the problem?

Most of the time it is caused by uneven lighting, bent paper, or the shooting angle is too oblique. First check whether the photo itself is blurry, and then decide whether to retake it. If there are still typos after reshooting, you need to manually correct them in the editor and check the key content word for word.

4. What is the difference between taking a photo of a page and using a scanner?

The scanner's light source and angle are fixed, and the paper is flattened, so the results are generally more stable than with a mobile phone. Taking pictures with a mobile phone is flexible and convenient, but it has higher requirements on light, flatness and framing. If the document is important and a scanner is available, use the scanner first.

📝 Summary

Scans do not necessarily require special equipment. As long as the photos are clear, flat, and complete, photos taken with mobile phones can also be used for OCR recognition. If you follow the self-check list before taking pictures and check each item after identification, you can significantly reduce rework. Multi-page documents are first merged into PDF, and then identified and checked page by page in OmniDoc. The result will be easier to check as a whole.

Open OmniDoc to recognize and translate scans