Design And Implementation Of Figure Photo Classification Software Based On Android

In the era of mobile Internet, smart phones have penetrated into every aspect of people’s social life, such as entertainment, shopping, banking etc. By the advantage of convenience and portability, phone camera has become one of the core functions of mobile phone. Though the advances in flash memory ha ve relieved the shackle of storage space, the problem of photos’ disorder has become increasingly prominent. Besides the classification by time and storage location, there is no solution that is able to handle the problem by now. As the most popular mobile operating system, Android has the highest urgency. Therefore, this paper designs and develops an intelligent figure photo classification software based on Android. Detailed work of the thesis is as follows:(1) Give a detailed introduction on Android platform and describe the software development environment and plug- ins used. In addition, core algorithms for face detection and face recognition are discussed especially.(2) Conduct a detailed requirement analysis and feasibility analysis according to users’ functional requirements for photo classification software. This paper proposes classification of figure photos as the core functional module and determines its functional process. At the same time, this paper proposes the album as auxiliary module which displays photos by person and determines its managerial function on single album and photo. In addition, the paper proposes client/server module as the solution against the classification of huge member of photos, and determines the reasonable way to insert it into core module.(3) Design all functional modules in detail, including photo management and classification entrance of home page, album functional module, photo classification module and client/server module. What’s more, this paper designs storage structure of the software.(4) Implement all functional modules on Android smart phone(Motorola ME525+ here) by Eclipse IDE. In order to improve the realtime performance, the software adopts JNI and NDK technologies to realize key algorithms of face detection and face recognition. Meanwhile, the software succeeds in fast reading of file by means of data format conversion. Performance testing is cond ucted, the result validates software’s real-time performance and effectiveness on figure photo classification.(5) Realize the uplink and downlink part of client/server module.