Handwriting Recognition Data Platform As A Service Based On OpenStack

In recent years, handwriting interaction is greatly promoted by the popularity oftouch-screen, and the value of handwriting recognition comes evident. Meanwhile, theprocess of industry information began to face the challenges of big data, including theoperation and maintenance of data, data resources re-allocation, data conversion of the valueto the business，etc.. It results in the industry’s cloud philosophy. Industries in business andacademia are facing such adjustments, the field of handwriting recognition industry, whichstudied in this paper, is also facing the opportunities and challenges of big data. Therefore,this research proposed an industry cloud solution for the field of handwriting recognition.In this paper, based on open source cloud computing platform OpenStack, we constructedthe Handwriting Recognition Data Platform as a Service (HRDPaaS). It’s going to solve theproblems currently encountered in handwriting recognition: a) it is difficult to achieve thehandwriting recognition on low-profile client (such as an embedded, Web-side); b) it requireslarge-storage and high-profile for personalized handwriting recognition; c) handwritingrecognition research needs to high-cost and large-number of sample sets; d) in the field ofhandwriting recognition, data resources is not balanced, and hardly data services are provided.HRDPaaS is specially presented for the field of handwriting recognition, handwrittendata as the core, providing handwriting recognition services, such as handwritten datastatistical service, handwriting sample scheduling service, data visualization services, etc.Meanwhile, with docking the platform-based resource monitoring system and the dataoperation and maintenance system, and using the flexibility of OpenStack-based scalable andload balancing solutions, we achieve deployment and scheduling the computing resources onthe platform.HRDPaaS achieves the following systems and services: a) Online handwritingrecognition; b) mass data storage of large number of writers contributes to personalizedhandwriting recognition research. c) providing a large number of sample scheduling; d)massive sample set implies natural semantics, it provides a basis for the transformation fromhandwritten sample data to semantics. Using this platform, we can integrate, optimize andrestructuring the ecological chain of handwriting interaction.