In Malaysia, signages in the Malay language are ubiquitous, be it in physical boards or other printed media. Since Malaysia is a well known foreign tourist destination, very often these signages may not be understandable to foreign tourists as these often come without English subtitles and tourists would be forced to consult any nearby passerby for a translation. In this project, optical character recognition technology was deployed within a mobile application to capture the images of these signs, apply optical character recognition to read these signs and finally translate it to English. With it, tourists or even young children with poor grasp of the Malay language can simply point their mobile phone cameras at a signage and expect an accurate translation in the English language. Additional features like text-to-speech and image archives were added to enrich the user experience in the app while optimizations to the translation engine such as a heuristic based dictionary and location based filtering were implemented to improve the translation time.

This system uses a client-server architecture with a connection module established to a cloud server. The frontend mobile application was developed with Flutter while Google Cloud Platform will be used as the backend platform. I hope that this application can break language barriers and make Malaysia a more welcoming destination for tourists.