Research on Interpretable Deep Learning-Driven Indoor Sports Visual Comfort Prediction
QIU Jinghan
SHI Ligang
Abstract：Sports visual comfort is a crucial factor influencing athletes'performance,health,and safety.Due to limitations in sample size and methodology,previous research on predicting sports visual comfort has been scarce,leading to a rather random and unstructured approach to the design of natural lighting environment in sports venues.Introducing deep learning algorithms and data-driven methods helps address complex issues such as human perception in sports and the interaction with the environment.In order to establish a high-accuracy and interpretable visual comfort prediction model,this study is based on the measurements of indoor sports space lighting environment,questionnaires on athletes'visual comfort,and physiological measurement techniques.Deep learning algorithms are employed to build a sports visual comfort evaluation model,and the contribution of each influencing factor to athletes'visual comfort is calculated.The results indicate that the deep learning models established in this study can predict the overall visual comfort of athletes,with the XGBoost model performing the best in predicting visual comfort.This model can reveal the complex relationships between various influencing factors and visual comfort,providing a basis for the future healthy design of sports venues.
Keywords：exerciserhealthy light environmentvisual comfortdeep learninginterpretability analysis
Publication Date：2025-02-27
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：11( 41-51 )