A Survey on Image-Based Approaches for Android Malware Detection: Toward Sustainable and Efficient Solutions


Sezer O., Doğru İ. A., Kılıç K.

Book of Abstracts of the International Conference on Sustainability – Transforming Higher Education for a Sustainable Future, Susana Leal,Cláudio Barradas,Ana Loureiro,Inês Messias,Sandra Oliveira, Editör, IPSantarém, Lisbon, ss.109, 2025

  • Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
  • Basım Tarihi: 2025
  • Yayınevi: IPSantarém
  • Basıldığı Şehir: Lisbon
  • Sayfa Sayıları: ss.109
  • Editörler: Susana Leal,Cláudio Barradas,Ana Loureiro,Inês Messias,Sandra Oliveira, Editör
  • Gazi Üniversitesi Adresli: Evet

Özet

The rapid proliferation of mobile device usage has introduced significant cybersecurity

threats. Among them, the Android operating system has become one of the primary targets

for malware developers due to its open-source nature and vast user base. While traditional

malware detection methods rely on static and dynamic analysis techniques, these often

require extensive preprocessing and expert-driven feature extraction and selection, which

can limit their efficiency and introduce performance bottlenecks. Furthermore,

conventional approaches may fall short when confronted with advanced and sophisticated

malware variants. In recent years, deep learning and image processing-based approaches

have emerged as innovative and effective alternatives for malware detection. These

methods not only enhance detection accuracy but also contribute to a sustainable and

scalable cybersecurity infrastructure through their automation capabilities. This survey

reviews the current literature on image-based methods for Android malware detection,

providing a detailed analysis of the applied techniques, their strengths, and their

limitations. In particular, approaches involving deep learning, convolutional neural

networks (CNNs), and other machine learning algorithms are comparatively evaluated. The

findings indicate that image-based analysis methods offer more reliable, comprehensive,

and effective detection than traditional techniques. Moreover, these approaches hold

significant promise for advancing sustainable digital security systems. Recommendations

for future research directions are also presented, along with a discussion of the potential

contributions to the academic body of knowledge in this domain.