DNN Architecture for High Performance Prediction on Natural Videos Loses Submodule's Ability to Learn Discrete-World Dataset.
Lana SinapayenAtsushi NodaPublished in: ALIFE (2019)
Keyphrases
- prediction accuracy
- video dataset
- sequence prediction
- video sequences
- prediction error
- prediction model
- human actions
- real time
- grey prediction model
- prediction algorithm
- software architecture
- management system
- training data
- video frames
- real world
- human beings
- trecvid multimedia event detection
- computation intensive
- space time interest points
- photo collections
- dynamic scenes
- distributed memory
- video database
- training process
- video clips
- human activities
- benchmark datasets
- neural network