On the evaluation of android malware detectors against code-obfuscation techniques

Area of Science:

• Computer Science
• Cybersecurity
• Software Engineering

Background:

• Android's market dominance makes it a prime target for malware.
• Malware developers employ code obfuscation to evade anti-malware detection.
• Existing research often analyzes individual or simple hybrid obfuscation techniques.

Purpose of the Study:

• To comprehensively evaluate anti-malware tools against complex hybrid code obfuscation techniques.
• To assess the impact of different obfuscation strategies on malware detection rates.
• To identify the effectiveness of inter-category obfuscations compared to simpler methods.

Main Methods:

• Evaluation of prominent anti-malware tools.
• Analysis of individual, simple hybrid, and inter-category code obfuscation techniques.
• Comparative assessment of detection rates across various obfuscation methods.

Main Results:

• Individual obfuscation techniques achieved nearly 100% detection for most tools.
• Category-wise obfuscation reduced detection rates for four out of ten tools.
• Inter-category hybridized obfuscations proved most effective at evading detection, with no tool achieving full detection.

Conclusions:

• Complex, inter-category hybrid obfuscations pose a significant threat to Android malware detection.
• Current anti-malware tools struggle against sophisticated obfuscation strategies.
• Further research is needed to develop robust detection mechanisms for advanced obfuscation techniques.