Abstract—The continuous increment of malware and its complexity motivated researchers to implement techniques to detect and classify it. Manual detection of malicious files is time consuming and shows poor results. Recently, Deep Convolution Neural Networks (DCNN) shows promising results in malware detection. DCNNs include large number of fully connected layers that are capable to deal with fast iterations of Android malware. Compared to the existing approach, DCNN shows high performance and accuracy in detecting different types of malwares. The proposed work combines Scale-Invariant Feature Transform (SIFT) and DCNN to detect malware features. Combining SIFT with DCNN allow higher accuracy of features classification and overcome the problem of single-feature extraction. The proposed method is compared to existing approaches to malware detection in terms of anticipated time and detection accuracy. The experimental results showed the significant enhancement offered by the proposed work in terms of accuracy and performance.
 
Keywords—malware, Deep Convolution Neural Networks (DCNN), Scale-Invariant Feature Transform (SIFT), color image transformation


Copyright © 2023 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.