| Android system occupy a dominant position in the smart phone market,which attracts many malicious developers to modify the Android application.Mobile devices contain various privacy information of users,especially with the popular ity of mobile payments,users download and use repackaged applications,it is easy to disclose personal information.In view of this situation,the research of program similarity detection based on Android system has the protection of user privacy information and practical value.This thesis mainly completed the following work:(1)This thesis is aimed at the similar application detection of Android system,analyzes the traditional detection technology,and proposes a static detection method to generate function call graph of the application.Decompiling by reverse analysis to get Smali source,and analyzing the source code generation function.Process the function call graph by filtering out the third-party library and related interfaces.(2)This thesis uses a two-dimensional adjacency matrix to represent the function call graph.By adding information such as the degree of the function node and the degree of ingress,the adjacency matrix is linearly transformed to obtain the corresponding feature vector.The similar ity between the feature vectors is calculated using the combination formula of the angle cosine and the Jaccard to determine the similarity of the Android application.(3)According to the designed detection system,a large number of different experiments are carried out,and the similarity is obtained to determine whether it is a repackaging application.Then,different types of applications on the current application market are tested to verify the accuracy and efficiency of the system detection.Experiments show that the similar ity detection system designed in this thesis can effectively analyze malicious repackaging application,with high accuracy and performance,and reduce the risk of repackaged applications invading smart phones. |
PDF Full Text Request