Selective Social-Interaction via Individual Importance for Fast Human Trajectory Prediction
This addresses computational efficiency in human trajectory prediction for robotics or surveillance applications, but appears incremental.
The paper tackles the problem of efficiently predicting human trajectories by selecting only important neighboring people, achieving competitive prediction accuracy while speeding up the process on the JRDB dataset.
This paper presents an architecture for selecting important neighboring people to predict the primary person's trajectory. To achieve effective neighboring people selection, we propose a people selection module called the Importance Estimator which outputs the importance of each neighboring person for predicting the primary person's future trajectory. To prevent gradients from being blocked by non-differentiable operations when sampling surrounding people based on their importance, we employ the Gumbel Softmax for training. Experiments conducted on the JRDB dataset show that our method speeds up the process with competitive prediction accuracy.