Research And Implementation Of Content Security Audit System For Digital Copier

In recent years,manufactures add features and new functions on their digital copiers to meet the demand of informational work way.In enterprises and institutions,although the digital copiers make work and life convenient,security issues on them have caused extensive concern,such as information leak,actions beyond authority,malicious behavior and so on.In order to enhance supervision of digital copiers and protect information security,regular organization of security inspection for digital copiers become more and more important.Therefore,it is of great significance to find out a proper way of digital copiers'content audit.According to this background,there are three major solutions for digital copiers auditing in the market:security-enhanced digital copiers with auditing capability,print content monitoring systems based on print protocols,and forensics systems based on content recovery for hard disks of digital copiers.However,there are some problems in the above schemes,such as high audit costs,incomplete audit objects and lack of behavior audit.Based on the above analysis,we focus on content audit and behavior audit,and implement an audit system for digital copiers'content,including storage images,logs and configurations.The specific work of this paper is divided into the following four parts:(1)As for audit content acquirement,we find that digital copiers have web systems and use this feature as an interface.We summarize the request parameters and URLs to obtain images,logs and configurations for seven mainstream brand digital copiers.Based on this,we utilize web crawler to automatically get content from digital copiers,which can enhance compatibility,help collect the content and make audit easier.(2)As for content audit,we use mature OCR tool to extract text content from the images,and extract sensitive content based on keyword matching technology.However,we find that the MMR format images in digital copiers can't be recognized by OCR tool and mainstream image process software.To solve this problem,we propose a MMR transformation method to transform the MMR images into TIFF images,and make the new images readable and recognizable.(3)As for behavior audit,we focus on operation event audit and users'behavior audit based on logs.In the process of operation event audit,we first analyze the original logs,combine multiple logs which represent a complete operation event,and generate log sequences.And then,we extract the key features of each event log sequence,and use agglomerative clustering algorithm to obtain the outliers,which will be reported to the auditors as suspicious events.In the process of users' behavior audit,we use a modified Apriori algorithm to make association analysis on a transaction dataset which consists of user identity,time and event category,and build a users' behavior model.The behaviors that can't match the model will be reported as abnormal behaviors.The scheme can effectively compensate for the lack of behavior audit in the existing schemes.(4)Based on the previous work,we implement an audit system for digital copiers' content in a modular way,which support more than a hundred models.The experiments show that the system is stable,and can get content accurately,which can meet the demand of digital copiers'content audit and behavior audit for enterprises and institutions.