STINet: Spatio-Temporal-Interactive Network for Pedestrian Detection and Trajectory Prediction
This work addresses a critical challenge for autonomous vehicles by improving pedestrian safety through enhanced detection and prediction, though it is incremental as it builds on existing two-stage and interaction modeling approaches.
The paper tackles the joint problem of pedestrian detection and trajectory prediction for autonomous driving by proposing STINet, an end-to-end two-stage network that integrates spatio-temporal modeling and object interactions, achieving a BEV detection AP of 80.73 and trajectory prediction ADE of 33.67cm on the Waymo Open Dataset, setting new state-of-the-art results.
Detecting pedestrians and predicting future trajectories for them are critical tasks for numerous applications, such as autonomous driving. Previous methods either treat the detection and prediction as separate tasks or simply add a trajectory regression head on top of a detector. In this work, we present a novel end-to-end two-stage network: Spatio-Temporal-Interactive Network (STINet). In addition to 3D geometry modeling of pedestrians, we model the temporal information for each of the pedestrians. To do so, our method predicts both current and past locations in the first stage, so that each pedestrian can be linked across frames and the comprehensive spatio-temporal information can be captured in the second stage. Also, we model the interaction among objects with an interaction graph, to gather the information among the neighboring objects. Comprehensive experiments on the Lyft Dataset and the recently released large-scale Waymo Open Dataset for both object detection and future trajectory prediction validate the effectiveness of the proposed method. For the Waymo Open Dataset, we achieve a bird-eyes-view (BEV) detection AP of 80.73 and trajectory prediction average displacement error (ADE) of 33.67cm for pedestrians, which establish the state-of-the-art for both tasks.