Image requirements

1. Image Quality Requirements

• Resolution: Recommended image size is 1920×1080 pixels (Full HD) to ensure that all textual information and the profile image are clearly visible and extractable.

• Focus and Sharpness: The image should be in sharp focus, with no blurred areas, to allow for precise OCR and face extraction. Autofocus should be enabled on the camera to ensure this.

• Lighting Conditions: Adequate, even lighting is essential. Avoid overly bright or dark environments, as well as glare or reflections that might obscure the text or photo. Users should avoid using flash if it creates reflections on the document surface.

• Color: It is recommended that the image is in color and not black and white or scanned.

​

2. Document Position and Margins

• Document Placement: It is recommended for the document to take up 50-80% of the image width, with small space around it (some margin should be visible). Too much empty space can lead to difficulties in cropping and recognizing the document. Avoid positioning the document too far from the camera. All the document edges should be in the image.

• Angle: The document should be captured at a straight angle, it shouldn't be tilted or slanted toward or away from you (Z-axis) more than 10 degrees. Extreme angles can distort the text and profile image, leading to inaccurate data extraction.

​

3. File Format and Size

• Supported Formats: The acceptable formats for document uploads are JPG and PNG. These formats provide compatibility across most devices and ensure high-quality image processing.

• File Size: Images should be between 100 KB and 10 MB in size, balancing quality and performance. Larger images may be automatically compressed, but excessively small images (<100 KB) may lack the necessary detail for accurate recognition.

​

4. Text and Profile Image Visibility

• Contrast: The document should be placed against a background that contrasts with the document color (e.g., a dark background for a light-colored document). This ensures the text and profile image are clearly distinguishable.

• Readability: All textual data, including the barcode, MRZ, and profile image, must be clearly visible and free from obstructions (e.g., fingers, glare, foreign objects).

  • Blur and Glare: Guide users to hold the camera parallel to the document, and ensure there is no direct lighting causing reflections.

  • Watermarks or Edits: Ensure the image is free of any third-party watermarks or edits, as these can reduce data extraction accuracy.

• Profile Image: The face image on the document must be well-lit and unobstructed for facial data extraction. Any shadows or light reflections on the profile image will reduce recognition accuracy.

​

5. Document Integrity and Validity

• Undamaged Document: The document should be free from physical damage, such as scratches, stains, or tears, that might interfere with data extraction.

• No Alterations: Uploaded images must be original, unedited, and not modified by any graphic software. Screenshots or photos from social media should be avoided, as they may introduce artifacts or compression issues.

​

7. API Integration and Error Handling

• Implement real-time feedback for users, providing them with warnings or suggestions (e.g., "Please ensure the document is in focus" or "Ensure the document is fully visible") during the upload process.

• Leverage an API for uploading and processing images that can dynamically handle variations in image size and provide precise error messages if the image does not meet the specified requirements.