Best practices for the digitization of documents

To ensure the best possible data extraction and OCR results, we recommend:

Recommendations

Documents should be scanned at least 300dpi.

What elements can affect image quality?

• Stains on original documents.
• Folds.
• Output format and depth when scanning.

What output formats do we recommend?

• PDF.
• TIFF.

However, Athento supports the formats listed in the following OCR Extraction.

What happens when the image quality is low?

Two processes are particularly affected:

• Classification based on textual expressions: As OCR cannot be extracted correctly, textual expressions are not extracted well either, so classification may fail.
• Metadata extraction: Metadata may not be extracted or may be extracted incorrectly.

Is processing a photo the same as processing a scanned image?

No, it is not the same. Processing photographs is much more complex, as camera-generated images can have a lot more complexity:

• Perspective distortion.
• Blur.
• Luminosity problems.

Can Athento correct image defects?

Athento has operations that can correct defects such as the following:

• Correct document orientation.
• Delete blank pages.
• Clean up OCR...etc.

However, corrections have limitations, and in many cases are not sufficient to obtain 100% optimal results.

You can also consult

What is Athento's success rate in data extraction?