Malware Image classification based on Mamba

Malicious code detection is one of the important research directions in the field of cybersecurity. Converting code into image information using convolutional neural networks (CNN) for malicious code detection has been applied. However, although converting the time series information in the code into two-dimensional information can fully utilize the computational efficiency of CNN, the impact of this two-dimensional processing is unknown. This article proposes a malicious code recognition method based on Mamba, which utilizes Mamba's advantages in processing long sequence data for malicious code recognition. This article evaluated the effectiveness and robustness of the model on the Malimg dataset, achieving a recognition accuracy rate of 98.56%.