Digital KYC For Bank With Python Development
About Project
The client, a leading Philippines financial services provider, desired to completely automate its KYC process of digitizing and extracting text from customers for perfect data accuracy and efficiency in document processing along with face scan and mapping with customer’s stored details.
Web
Platform/OSFinance
Category
Brief
We automated the data extraction of over 65,000 customer accounts. We also accurately verified the customer's selfie by running a check with the customer's stored photo. Brainvire combined OCR technology with Open Source solution for data accuracy and improved document processing efficiency.
Highlights
Using Backend API, we equipped the client to seamlessly scan documents over 65K customer accounts and extract the text accurately. To make the KYC process entirely digital, we enabled the Philippine Bank to verify the customer’s identity by simply scanning the customer’s face with a selfie. We also met the client’s requirement of minimum 95% accuracy rate for printed and 85% for handwritten text from scanned customer documentation.
Case StudyFeatures
Capability To Scan Minimum 65,000 Documents:
The client was able to scan at least 65K customer accounts documents with Brainvire’s implementation of Lambda function and a robust server architecture for a smooth data capture.
Selfie Mapping Of Customer’s Photo:
The client simplified the KYC verification process by accurately mapping customer selfie with actual photo ID
95% Scanned Document Accuracy:
Using Google Vision APIs, we helped the client unlock unmatched accuracy of minimum 95% for printed and 85% for handwritten text from documents.
Image Comparison Without Database Storage:
To successfully match the customer’s image without storing it in the database, we used the NumPy module, and stored the received base-64 encoded image in a temporary variable only till the comparison process
Tech Stack
Python
Google Vision
