基于表征的安卓重打包检测方案
首发时间:2019-01-17
摘要:目前的安卓重打包检测方法多采用代码层面的特征进行相似度的计算,这种方法存在无法有效应对代码混淆的缺点,同时无法有效检测出钓鱼形式的重打包应用。本文提出了一种通过提取应用一些典型的表征信息,例如名称、应用图标等,借助多维度鲁棒的相似度算法,进行快速有效的重打包检测模型。该模型在从主流应用市场获取的应用数据集上达到了97的F1分数。
For information in English, please click here
Android repackaging detection based on characteristics
Abstract:The current Android repackage matching method mostly compute similarity based on feature extracted from reversed code. This method is not able to solve code confusion effectively, and also not able to detect fishing Android app appearing on market frequently. Based on features extracted from app\'s characteristics, such as app name, package name and icon, fast and robust repackaging detection is doable, and achieves an F1 score of 97 on the dataset crawled from mainstream markets.
Keywords: Android repackaging detection characteristics
基金:
引用
No.****
同行评议
勘误表
基于表征的安卓重打包检测方案
评论
全部评论0/1000