engleski

Identifying software modules in executable files

This thesis describes a platform-independent system designed to be capable of automated detection of malicious programs, re-used parts of code, statically linked third party code and vulnerable functions in executable code. The core of the system is an algorithm for generating function signatures that captures functionality, while being resistant to non-functional changes. The system was implemented for the Common Language Infrastructure platform, and evaluated as a library function recognition system and a malware detection system. The evaluation results show that the system has potential to be used in solving real-world problems.

Function signatures, library function recognition, malware detection

nije evidentirano