Image Data Requirements
Our document processing system is engineered to accommodate a broad spectrum of image qualities, captured under various lighting conditions and orientations. To ensure the highest level of accuracy in our anti-spoofing measures, we recommend adhering to the following guidelines:
Supported Image Size and Compression
- Image File Size: Images must not exceed a file size of 4.4 MB to maintain efficient performance and resource utilization.
- Minimum Resolution Requirement: The API requires images to be at least HD quality, with minimum dimensions of 1,280x720 pixels and maximum dimension 10,000x10,000 pixels
- Compression Guidelines: To preserve image quality, images should be compressed using a single pass at a JPEG quality of 70 or higher. Uncompressed images are preferred to avoid the loss of quality from multiple compressions. Repeated compression should be avoided.
Supported Image format
- JPEG Files (*.jpeg, *.jpg, *.jpe)
- Portable Network Graphics -*.png
- WebP - *.webp
- Portable Document Format- *.pdf
- High Efficiency Image File Container- *.HEIC
- Portable Image Format ( *.pbm, *.pgm, *.ppm, *.pxm, *.pnm)
Image Quality Requirements
Below are the requirements for the quality of document images captured by any device, which are necessary for successful image processing:
- Document Borders: Ensure that the image includes a border around the document that is at least 5% of the document's width. Images that are too tightly cropped may present challenges in accurate evaluation.
- Image Clarity: Avoid blurring, which can occur from motion during capture, as it may result in the rejection of legitimate documents. Similarly, image smoothing and texture filtering could increase the likelihood of approving spoofed images.
- Lighting and Glare: Direct sunlight or sharp overhead lighting can create glare, increasing the risk of document rejection. Optimal lighting conditions without artificial coloration are crucial for accurate processing.
- Perspective Distortion: Images that exhibit significant perspective distortion cannot be processed accurately. Perspective distortion occurs when an object or scene is photographed from an angle rather than straight on, leading to a skewed or stretched appearance in the image.
Updated 5 months ago