Research Article
Automated Script Recognition Based on Human-Computer Interaction Features of Mouse Movement Trajectory
@INPROCEEDINGS{10.4108/eai.22-5-2026.2365097, author={Chenyang Wu}, title={Automated Script Recognition Based on Human-Computer Interaction Features of Mouse Movement Trajectory}, proceedings={Proceedings of the 4th International Conference on Image, Algorithms, and Artificial Intelligence, ICIAAI 2026, 22-24 May 2026, Singapore, Singapore}, publisher={EAI}, proceedings_a={ICIAAI}, year={2026}, month={8}, keywords={Human-computer interaction Mouse trajectory Human-computer recognition Machine learning}, doi={10.4108/eai.22-5-2026.2365097} }- Chenyang Wu
Year: 2026
Automated Script Recognition Based on Human-Computer Interaction Features of Mouse Movement Trajectory
ICIAAI
EAI
DOI: 10.4108/eai.22-5-2026.2365097
Abstract
As human-computer interaction scenarios become increasingly diverse, human-computer recognition technology has become crucial. While ensuring that human-machine recognition technology cannot be cracked, user experience should also be taken into account.This paper extracts the kinematic features and human-computer interaction features of mouse trajectories.Having done 10 cross-validations on a training sample containing 3000 mouse trajectories, it was then inputted as the optimal parameters and the training sample into an Extreme Gradient Boosting (XGBoost) binary classification model to be trained. In the last model, human behavior recognition rate is 97.974 percent and machine behavior recognition rate is 98.012 percent, and the F1 score of had been obtained to be 98.04784. This paper employs machine learning to recommend a new strategy of human-machine recognition.
