About this document

Pedestrian Crossing Direction Prediction by krishnasamykiruthiga is a document available to read on EtoBox.

This study presents a novel framework for predicting pedestrian crossing directions at intersections to enhance safety, utilizing keypoints and trajectory data from CCTV footage. The framework employs Transformer-based models and Graph Convolutional Networks (GCNs), achieving an accuracy of 94.10% and an F1-Score of 92.35%. By standardizing spatial features across varying intersection geometries and camera perspectives, the model aims to reduce pedestrian-vehicle conflicts and can be integrated into traffic

Author

krishnasamykiruthiga

Language

EN