Skip to content

Image Quality Consideration

This document describes the image quality considerations for completing the processing in the backend, and the recommendation to get the best possible validation. These images are the ones that should go directly to the API call, and they refer to all the images processed front, and back.

Image Format

The product supports any type of image format but we recommends the following extensions JPEG, PNG, and BMP, for getting the best result possible.

Minimum Requirements

Document orientation

The product suport a maximun rotation of ± 20º. The best rotation recommended is ±10º. The document should always be horizontal compared to the image. The following examples show a good orientation and a bad one.

OK Validation examples

Ok and KO Validation examples

Rotations and extreme perspective distortion (right picture) are corrected before document analysis.

Correct perspective example

Image Document Proportion

Documents width and height must be at least 40% and a maximum of 90% of the width and height of the image. The document area should have a minimum resolution of 0.5MPx and a maximum of 30MPx. However, higher resolutions are recommended to reach optimal extraction than lower resolutions.

Image dimension example

Image Recomendations

For getting the best validation possible we recommend a series of points. Which refer to the orientation of the image, the quality of the image, and the capture conditions.

Images orientation

  • The document always come horizontal with a rotation no bigger than ± 10º.
  • The document should represent a minimum of 60% of the image.
  • The tilt angle of the document should be bigger than ± 5º.

The example below meets the requirements mentioned above.

Recomendation Orientation example

Images quality

The document should avoid any brightness or obstruction of the content text, MRZ, or any barcode. It will generate errors in the validation or in the OCR extraction.

Obstruction of teh MRZ example

The image should be as clear as possible, avoiding blurred images.

Blur Image example

Capture conditions

The capture general condition should have proper lighting. It helps the OCR extraction and validation.

Bad light Image example

The capture should not have the presence of color dominants. This makes the algorithms not perform as best as possible.

Color dominat Image example