Spike-FlowNet: Event-Based Optical Flow Estimation with Energy-Efficient Hybrid Neural Networks.
Chankyu LeeAdarsh KostaAlex Zihao ZhuKenneth ChaneyKostas DaniilidisKaushik RoyPublished in: ECCV (29) (2020)
Keyphrases
- energy efficient
- optical flow estimation
- hybrid neural
- optical flow
- wireless sensor networks
- energy consumption
- neural network
- flow field
- sensor networks
- learning rules
- stereo matching
- motion model
- base station
- data transmission
- hopfield neural network
- motion field
- computer networks
- multiscale
- coarse to fine
- fuzzy systems
- sensor nodes
- energy efficiency
- velocity field
- routing protocol
- artificial neural networks
- feature matching
- image processing
- routing algorithm
- network structure
- biologically inspired
- motion segmentation
- power consumption
- feature points
- higher order
- motion estimation
- computational complexity
- three dimensional
- genetic algorithm