Project Synopsis
A startup from the US wanted to roll out a functionality where customers could authenticate themselves by uploading their document ids using a mobile app. The user could click a photograph of a document using his phone camera and upload it for processing. The system was expected to process the image, validate if the uploaded document was authentic or fraudulent, extract the document id, date of birth, address etc.
Further, to ensure that the user was not using a stolen document, user had to upload a selfie of himself. This image of the user had to be compared with the image extracted from the document to check if it was the same user.
The developed platform was expected to have the following features:
- Validate authenticity of the user’s document
- Extract user information from the document
- Compare user’s selfie with the image extracted from the document
Our Solution
We designed and developed the complete solution for mobile platform. The delivered solution had the following features:
- Support for multiple document ids (passports, driver’s license etc.)
- Algorithms for image enhancements and clean up before processing
- OCR recognition and validation of extracted text
- Validating the document’s authenticity and identifying fraudulent documents
- Capturing user’s face and matching it with the picture in the document
Project Highlights
- Handled perspective correction and de-skewing of skewed documents
- Used live detection to ensure that user takes a selfie (and does not take a picture of an image) for the face recognition
- Complete document authentication less than 15 seconds
- Ability to validate the address provided in the document
