AMALGAN: Image-Based Android Malware Classification Using Generative Adversarial Network

The Android malware detection process requires analysing numerous files to ensure system security. Malware can also be embedded in media files and images. Android malware leverages the Android platform to propagate malicious payloads through image files created by malware developers. The proposed Android malware detection using the generative adversarial network (AMALGAN) approach employs image data to identify and classify Android malware. Unlike traditional uses of generative adversarial networks (GANs) for data generation or augmentation, the proposed approach exploits GANs innovatively for malware detection. In this work, the two modules of GAN, the generator and discriminator, are fine-tuned to identify and classify malware. An image-based malware dataset is used for training and validation. The GAN functions as an auxiliary classifier, while a hybrid malware analysis technique is applied to extract relevant features. Standard evaluation metrics are employed to assess the effectiveness of the proposed AMALGAN model. The results demonstrate that AMALGAN achieves 95.24% accuracy, 94.31% precision, 96.12% recall, and a 94.11% F1-score. These results confirm that the proposed approach outperforms several state-of-the-art methods in terms of accuracy, precision, recall, and F1-score.