Mealin S, Cleghern Z, Foster M, Bozkurt A, Roberts DL. Predicting Guide Dog Temperament Evaluation Outcomes Using Raw ECG Signals. ACI '19: Proceedings of the Sixth International Conference on Animal-Computer Interaction. 2020; Article No. 2:1-12. doi:10.1145/3371049.3371053.
This study explored the use of raw electrocardiography (ECG) signals to predict guide dog temperament evaluation outcomes. At Guiding Eyes for the Blind, 29 behavioral scores were predicted using ECG data collected during 11 different tasks with an accuracy of approximately 92%. The researchers utilized both a prototype and optimized network to assess these predictions. Findings suggest that the predictive features may be task-independent, as each task was capable of predicting any of the scores.
Commentaires