Control flow obfuscation based protection method for Android applications

2017 ◽  
Vol 14 (11) ◽  
pp. 247-259 ◽  
Author(s):  
Yong Peng ◽  
Guanyu Su ◽  
Bin Tian ◽  
Maohua Sun ◽  
Qi Li
2016 ◽  
Vol 61 ◽  
pp. 72-93 ◽  
Author(s):  
Vivek Balachandran ◽  
Sufatrio ◽  
Darell J.J. Tan ◽  
Vrizlynn L.L. Thing

Sensors ◽  
2020 ◽  
Vol 20 (15) ◽  
pp. 4201
Author(s):  
Yu-an Tan ◽  
Shuo Feng ◽  
Xiaochun Cheng ◽  
Yuanzhang Li ◽  
Jun Zheng

Information leaks can occur through many Android applications, including unauthorized access to sensors data. Hooking is an important technique for protecting Android applications and add security features to them even without its source code. Various hooking frameworks are developed to intercept events and process their own specific events. The hooking tools for Java methods are varied, however, the native hook has few methods. Besides, the commonly used Android hook frameworks cannot meet the requirement of hooking the native methods in shared libraries on non-root devices. Even though some approaches are able to hook these methods, they have limitations or are complicated to implement. In the paper, a feasible hooking approach for Android native methods is proposed and implemented, which does not need any modifications to both the Android framework and app’s code. In this approach, the method’s reference address is modified and control flow is redirected. Beyond that, this study combines this approach with VirtualXposed which aims to run it without root privileges. This hooking framework can be used to enforce security policies and monitor sensitive methods in shared objects. The evaluation of the scheme demonstrates its capability to perform hook operation without a significant runtime performance overhead on real devices and it is compatible and functional for the native hook.


Author(s):  
Manokaran Newlin Rajkumar ◽  
Varadhan Venkatesa Kumar ◽  
Ramachandhiran Vijayabhasker

This modern era of technological advancements facilitates the people to possess high-end smart phones with incredible features. With the increase in the number of mobile applications, we are witnessing the humongous increase in the malicious applications. Since most of the Android applications are available open source and used frequently in the smart phones, they are more vulnerable. Statistical and dynamical-based malware detection approaches are available to verify whether the mobile application is a genuine one, but only to a certain extent, as the level of mobile application scanning done by the said approaches are in general routine or a common, pre-specified pattern using the structure of control flow, information flow, API call, etc. A hybrid method based on deep learning methodology is proposed to identify the malicious applications in Android-based smart phones in this chapter, which embeds the possible merits of both the statistical-based malware detection approaches and dynamical-based malware detection approaches and minimizes the demerits of them.


Sign in / Sign up

Export Citation Format

Share Document