SpikeMS: Deep Spiking Neural Network for Motion Segmentation.
Chethan M. ParameshwaraSimin LiCornelia FermüllerNitin J. SanketMatthew EvanusaYiannis AloimonosPublished in: CoRR (2021)
Keyphrases
- motion segmentation
- spiking neural networks
- biologically inspired
- cerebellar model
- motion estimation
- image sequences
- motion model
- moving objects
- liquid state machine
- optical flow
- feed forward
- learning rules
- missing data
- multibody
- epipolar geometry estimation
- biologically plausible
- model selection
- artificial neural networks
- motor control
- spectral clustering
- neural network
- motion field
- training algorithm
- receptive fields
- similarity measure
- motion compensation
- computational model
- pose estimation
- multiscale
- three dimensional
- computer vision