东南大学学报(自然科学版)
東南大學學報(自然科學版)
동남대학학보(자연과학판)
Journal of Southeast University (Natural Science Edition)
2015年
6期
1046-1050
,共5页
王研昊%马媛媛%杨明%罗军舟
王研昊%馬媛媛%楊明%囉軍舟
왕연호%마원원%양명%라군주
Android 系统%智能终端识别%隐性标识符%指纹匹配算法
Android 繫統%智能終耑識彆%隱性標識符%指紋匹配算法
Android 계통%지능종단식별%은성표식부%지문필배산법
Android system%device identification%implicit identifiers%fingerprint matching algo-rithm
针对现有 Android 智能终端识别直接利用 IMEI,Android_ID 等显性标识符存在依赖敏感权限和易失效的问题,提出了一种基于隐性标识符组合的零权限设备识别方法.通过调用系统API 并执行 Linux Shell 命令,从物理层、应用层以及用户层3个层次上获取设备型号、屏幕信息、内核编译信息、User Agent、系统语言、字体大小、字体列表、用户安装程序列表等8个隐性标识符,并将其组合形成设备指纹.然后,提出了指纹精确匹配算法和变化指纹关联匹配算法,用于仅1个隐性标识符发生变化且变化前后相似程度大于0.85时的设备关联识别.通过采集真实用户数据进行实验验证,结果表明所提算法的设备识别准确率达到94.52%.
針對現有 Android 智能終耑識彆直接利用 IMEI,Android_ID 等顯性標識符存在依賴敏感權限和易失效的問題,提齣瞭一種基于隱性標識符組閤的零權限設備識彆方法.通過調用繫統API 併執行 Linux Shell 命令,從物理層、應用層以及用戶層3箇層次上穫取設備型號、屏幕信息、內覈編譯信息、User Agent、繫統語言、字體大小、字體列錶、用戶安裝程序列錶等8箇隱性標識符,併將其組閤形成設備指紋.然後,提齣瞭指紋精確匹配算法和變化指紋關聯匹配算法,用于僅1箇隱性標識符髮生變化且變化前後相似程度大于0.85時的設備關聯識彆.通過採集真實用戶數據進行實驗驗證,結果錶明所提算法的設備識彆準確率達到94.52%.
침대현유 Android 지능종단식별직접이용 IMEI,Android_ID 등현성표식부존재의뢰민감권한화역실효적문제,제출료일충기우은성표식부조합적령권한설비식별방법.통과조용계통API 병집행 Linux Shell 명령,종물리층、응용층이급용호층3개층차상획취설비형호、병막신식、내핵편역신식、User Agent、계통어언、자체대소、자체렬표、용호안장정서렬표등8개은성표식부,병장기조합형성설비지문.연후,제출료지문정학필배산법화변화지문관련필배산법,용우부1개은성표식부발생변화차변화전후상사정도대우0.85시적설비관련식별.통과채집진실용호수거진행실험험증,결과표명소제산법적설비식별준학솔체도94.52%.
In the current technologies of Android device identification,explicit identifiers are usually adopted,such as IMEI (international mobile equipment identity)and Android_ID,which depend on sensitive permissions and are liable to fail.To solve these problems,a zero permission device identification method based on the implicit identifier group is proposed.From the physic layer,the application layer and the user layer,eight implicit identifiers including device type,screen informa-tion,kernel compile information,user agent,system language,font size,font list and user installed package list are obtained by calling system API (application programming interface)and executing Linux Shell commands.These eight implicit identifiers are combined to form the device fingerprint. Then,the corresponding fingerprint exact matching algorithm and the changing fingerprint associated matching algorithm are proposed,which can be applied to device related identification when only one implicit identifier changes and the similarity ratio is above 0.85 before and after the change.Fi-nally,the experimental verification is carried out based on the real user data,and the results show that the device identification accuracy rate of the proposed algorithm can reach 94.52%.