使用生成对抗网络进行笔迹特征分析

Nov, 2016

使用生成对抗网络进行笔迹特征分析

Handwriting Profiling using Generative Adversarial Networks

Arna Ghosh, Biswarup Bhattacharya, Somnath Basu Roy Chowdhury

TL;DR使用改进的 DCGAN 架构和强化学习技术，本文提出一种智能系统，旨在学习实体的手写字体，可以用于鉴别伪造文件、签名验证、计算机生成艺术和文件数字化等领域。初步实现结果表明，该算法在 MNIST 数据集上表现优异。

Abstract

handwriting is a skill learned by humans from a very early age. The ability to develop one's own unique handwriting as well as mimic another person's →