论文标题
Mailleak：使用转移学习的混淆式刺激性角色提取
MailLeak: Obfuscation-Robust Character Extraction Using Transfer Learning
论文作者
论文摘要
以下工作提出了一种新的算法，用于从混淆的图像中识别角色。提出的方法是对当前邮政服务的潜在威胁的一个示例。本文均分析了给定算法的效率，并提出了预防这种威胁发生的对策。
The following work presents a new algorithm for character recognition from obfuscated images. The presented method is an example of a potential threat to current postal services. This paper both analyzes the efficiency of the given algorithm and suggests countermeasures to prevent such threats from occurring.
