Semantic-STGCNN is a research project developed as part of a master's thesis titled "Semantic Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction." The project aims to explore the application of graph convolutional neural networks (GCNNs) in predicting human trajectories in urban environments. By incorporating semantic information and leveraging spatio-temporal graph structures, the Semantic-STGCNN model offers a novel approach to address the challenges of trajectory prediction.

The Visualization script will allow you to test the Semantic-STGCNN model on a SDD dataset (https://paperswithcode.com/sota/trajectory-prediction-on-stanford-drone). This model Ranks 8th in the overall table with metrics of 10.93 picels for ADE and 18.44 pixels for FDE. There is an interactive ipwidget that you can play with at the end to test on single pedestrian trajectories.