Telling Human and Machine Handwriting Apart

TL;DR

This paper presents a highly accurate method for distinguishing human handwriting from machine-generated handwriting across multiple datasets and synthesizers, enhancing security verification systems.

Contribution

The study introduces a shallow recurrent neural network that effectively detects human versus machine handwriting using trajectory data, with strong performance even in few-shot and out-of-domain scenarios.

Findings

Achieves 98.3% AUC and 1.4% EER across datasets

Performs well with only 10% training data in few-shot settings

Maintains competitive results in out-of-domain tests

Abstract

Handwriting movements can be leveraged as a unique form of behavioral biometrics, to verify whether a real user is operating a device or application. This task can be framed as a reverse Turing test in which a computer has to detect if an input instance has been generated by a human or artificially. To tackle this task, we study ten public datasets of handwritten symbols (isolated characters, digits, gestures, pointing traces, and signatures) that are artificially reproduced using seven different synthesizers, including, among others, the Kinematic Theory (Sigma h model), generative adversarial networks, Transformers, and Diffusion models. We train a shallow recurrent neural network that achieves excellent performance (98.3 percent Area Under the ROC Curve (AUC) score and 1.4 percent equal error rate on average across all synthesizers and datasets) using nonfeaturized trajectory data as input. In few-shot settings, we show that our classifier achieves such an excellent performance when trained on just 10 percent of the data, as evaluated on the remaining 90% of the data as a test set. We further challenge our classifier in out-of-domain settings, and observe very competitive results as well. Our work has implications for computerized systems that need to verify human presence, and adds an additional layer of security to keep attackers at bay.