Multi-Head Attention for Multi-Modal Joint Vehicle Motion Forecasting.
Jean MercatThomas GillesNicole El ZoghbyGuillaume SandouDominique BeauvoisGuillermo Pita GilPublished in: ICRA (2020)
Keyphrases
- multi modal
- humanoid robot
- kinematic model
- multi modality
- image sequences
- audio visual
- optical flow
- space time
- short term
- spatial and temporal
- cross modal
- semantic concepts
- human motion
- image annotation
- high level
- fully automatic
- eye movements
- focus of attention
- high dimensional
- feature space
- similarity measure
- fusing multiple