Implementing Arabic Handwritten Recognition Approach using Cloud Computing and Google APIs on a mobile application

Abstract

Cloud computing application for automatic recognition of Arabic handwritten is needed nowadays and of great importance especially when implemented on a mobile application. It is a type of applications that includes many challenging aspects. This paper introduces for the first time a cloud based mobile app that applies Arabic handwritten recognition for translation purposes and finding locations on Google maps. Proposing such a service for non-Arabic speakers is of great importance especially while visiting Arabic speaking countries. Our approach is the first to build a mobile app based on cloud computing that proposes a multi-phase hybrid classifier for Arabic Handwritten text recognition. Google Maps and Google Translate APIs are applied on the recognized text as part of the introduced cloud computing application. The recognition part of the proposed approach is a multi-stage classifier introduced to cope with big database and high computation complexities. The experiment applied on our approach shows better results of our Arabic handwritten recognition when compared with similar approaches.