EV-MGRFlowNet: Motion-Guided Recurrent Network for Unsupervised Event-based Optical Flow with Hybrid Motion-Compensation Loss.
Hao ZhuangXinjie HuangKuanxu HouDelei KongChenming HuZheng FangPublished in: CoRR (2023)
Keyphrases
- motion compensation
- motion field
- optical flow
- motion estimation
- recurrent networks
- motion compensated
- motion vectors
- block matching
- motion model
- video coding
- optical flow estimation
- video compression
- image sequences
- motion estimation algorithm
- flow field
- motion segmentation
- motion parameters
- video super resolution
- recurrent neural networks
- camera motion
- feed forward
- motion estimates
- consecutive frames
- ego motion
- computer vision
- neural network
- inter frame
- reference frame
- moving objects
- velocity field
- block matching algorithm
- video codec
- vector field
- rigid motion
- motion compensated prediction
- beating heart
- multi frame
- video sequences
- moving camera
- super resolution
- phase correlation
- global motion
- error propagation
- dense motion field
- macroblock