Toward Better Pedestrian Trajectory Predictions: The Role of Density and Time-to-Collision in Hybrid Deep-Learning Algorithms.
Raphael KorbmacherAntoine TordeuxPublished in: Sensors (2024)
Keyphrases
- learning algorithm
- deep architectures
- active learning
- training data
- object detection
- machine learning
- collision free
- trajectory data
- efficient learning
- hybrid learning
- back propagation
- machine learning algorithms
- learning problems
- boolean functions
- cellular automata
- collision avoidance
- deep learning
- learning process
- computer vision