Reading handwritten customer forms with Cloud Vision API

The challenge

MPM needed to extract data from customer ID card, family card, and driving licence images and various forms to inform its marketing strategy. Much of the source information was handwritten and previously had to be retyped by hand, so it needed a tool that could read both typed and handwritten text.

What we did

• Proposed and implemented Google Cloud Vision API for AI and machine-learning image processing
• Used Vision API text and handwriting detection, which supports the languages on customer forms, including Bahasa Indonesia
• Delivered it as a data-capture capability that extracts information from ID and family registration cards

The results

• Stronger ability to collect and extract data from multiple sources
• More data available to build a better marketing strategy