A GAN-EfficientNet-Based Traceability Method for Malicious Code Variant Families
作者机构：School of Computer and Control EngineeringNortheast Forestry UniversityHarbin150040China
出 版 物：《Computers, Materials & Continua》 (计算机、材料和连续体（英文）)
年 卷 期：2024年第80卷第7期
页 面：801-818页
核心收录：
学科分类：0810[工学-信息与通信工程] 1205[管理学-图书情报与档案管理] 0839[工学-网络空间安全] 08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）]
基 金：support this work is the Key Research and Development Program of Heilongjiang Province specifically Grant Number 2023ZX02C10
主 题：Malicious code variant traceability feature reuse lightweight neural networks code visualization attention mechanism
摘 要：Due to the diversity and unpredictability of changes in malicious code,studying the traceability of variant families remains *** this paper,we propose a GAN-EfficientNetV2-based method for tracing families of malicious code *** method leverages the similarity in layouts and textures between images of malicious code variants from the same source and their original family of malicious code *** method includes a lightweight classifier and a *** classifier utilizes the enhanced EfficientNetV2 to categorize malicious code images and can be easily deployed on mobile,embedded,and other *** simulator utilizes an enhanced generative adversarial network to simulate different variants of malicious code and generates datasets to validate the model’s *** process helps identify model vulnerabilities and security risks,facilitating model enhancement and *** classifier achieves 98.61%and 97.59%accuracy on the MMCC dataset and Malevis dataset,*** simulator’s generated image of malicious code variants has an FID value of 155.44 and an IS value of 1.72±*** classifier’s accuracy for tracing the family of malicious code variants is as high as 90.29%,surpassing that of mainstream neural network *** meets the current demand for high generalization and anti-obfuscation abilities in malicious code classification models due to the rapid evolution of malicious code.
维普期刊数据库