基于卷积神经网络的恶意代码分类
首发时间:2016-12-05
摘要:提出一种基于恶意代码图像的恶意代码分类的检测方法,通过结合计算机视觉分析技术与恶意代码变种检测技术,将恶意代码映射为无压缩灰阶图片,基于属于相同族的恶意代码在恶意代码图像上比不同族的具有更明显的相似性假设,使用计算机视觉的卷积神经网络分类算法,构造恶意代码分类网络结构,在此基础上,实现恶意代码分类系统的训练与预测。并通过对25种恶意代码样本数据集的分析,训练和检测,完成了对该系统的实验验证。实验结果表明,基于上述方法分类方法具有检测速度快、识别率高等特点,并且对恶意代码变种具有较好的辨别能力。
For information in English, please click here
Malware Detection Based On Convolutional Neural Network
Abstract:A malware detection method based on malware images is proposed. By combining computer vision analysis technology and malware variants detection technology, the malware is mapped to non-compressed gray-level pictures. Based on the malware belonging to the same family in the malware image The classification of malware classification system is constructed based on the convolutional neural network classification algorithm of computer vision. On this basis, the training and prediction of malware classification system is realized. And through 25 kinds of malware sample dataset analysis, training and testing, completed the experimental verification of the system. The experimental results show that the classification method based on the above method has the characteristics of fast detection rate and high recognition rate, and has better discrimination ability against malware variants.
Keywords: Malware Malware Image Computer Vision Convolutional Neural Network
基金:
论文图表:
引用
No.4710789117097614****
同行评议
共计0人参与
勘误表
基于卷积神经网络的恶意代码分类
评论
全部评论0/1000