Virus-MNIST
收藏资源简介:
Virus-MNIST是一个用于恶意软件检测的图像分类数据集,由人民泰克公司创建。该数据集包含约50,000个病毒样本,涵盖9个计算机病毒家族和一个良性集合。数据集通过将可移植执行文件(PE)的前1024字节转换为32x32像素的灰度图像来创建,模拟了MNIST手写数据集的格式。创建过程中使用了KMeans聚类算法来识别病毒家族。Virus-MNIST主要应用于深度学习领域,旨在通过图像识别技术提高恶意软件的检测准确率,解决传统基于签名的检测方法的局限性。
Virus-MNIST is an image classification dataset for malware detection, created by PeopleTec Inc. This dataset contains approximately 50,000 malware samples, covering 9 computer virus families and one benign sample set. It is constructed by converting the first 1024 bytes of Portable Executable (PE) files into 32×32 grayscale images, mimicking the format of the MNIST handwritten digit dataset. The K-Means clustering algorithm was utilized during the dataset creation to identify virus families. Primarily applied in the deep learning field, Virus-MNIST aims to improve malware detection accuracy via image recognition technologies, addressing the limitations of traditional signature-based detection methods.




