.png){kind=icon}
返回
Multiclass malware classification via first- and second-order texture statistics
.png){kind=icon}
PRE
AI
DOI：10.1016/j.cose.2020.101895.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
The generally increasing volume of malware poses a challenge to the predominantly used static or dynamic analysis, which requires complex disassembly or time-intensive execution. Furthermore, high fea ture dimensionality and feature extraction costs per instance of malware increase overhead. The efficient classification of obfuscated malware, particularly for imbalanced classes, is a major challenge. This paper presents a visualization approach for malware classification to fill the gaps. Motivated by the visual similarity among malware from the same family, this paper proposes binary texture analysis over greyscale images created directly from their malware executables. The technique derives a novel combination of first-order and grey-level co-occurrence matrix (GLCM)-based second-order statistical texture features over the visualized malware. Using ensemble learning, the higher F1 score and accuracy obtained us ing Malimg (a benchmark Windows malware dataset) relative to those of state-of-the-art techniques and Windows executables first submitted to VirusTotal from 2018 to 2019 indicate higher efficiency and reliability of the proposed technique for malware classification. Additionally, the proposed technique is robust to obfuscation methods (e.g., packing, code relocation, and encryption). Indeed, the technique uses relatively fewer features extracted without disassembly or code execution in less time, thereby significantly enhancing the scalability of classifying a large-scale malware corpus. (c) 2020 Elsevier Ltd. All rights reserved.
Keyword:
Code obfuscation
Malware visualization
Binary texture analysis
GLCM
Ensemble learning
Malware classification
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
C
IF:
5.4
论文数:
4.6K
被引数:
1.4W
机构
引用论文
Early-stage malware prediction using recurrent neural networks基于递归神经网络的早期恶意软件预测.png)
COMPUTERS & SECURITY
OAAI
COMPUTERS & SECURITYIF5.4
OAAI分享收藏Improvement of malware detection and classification using API call sequence alignment and visualization使用API调用序列比对和可视化改进恶意软件检测和分类
PREAI
分享收藏Nanocrystalline silicon films as multifunctional material for optoelectronic and photovoltaic applications纳米硅薄膜作为光电和光伏应用的多功能材料
PREAI
分享收藏
查看更多内容