A Constrained Deformable Convolutional Network for Efficient Single Image Dynamic Scene Blind Deblurring with Spatially-Variant Motion Blur Kernels Estimation.
Shu TangYang WuHongxing QinXianzhong XieShuli YangJing WangPublished in: CoRR (2022)
Keyphrases
- blur kernel
- motion blur
- dynamic scenes
- single image
- blurred images
- spatially varying
- camera motion
- super resolution
- estimation algorithm
- video camera
- blind deconvolution
- image restoration
- point spread function
- space time
- image deblurring
- multi view
- frame rate
- motion model
- motion segmentation
- multiple images
- video sequences
- total variation
- natural images
- denoising
- shape from shading
- motion estimation
- high resolution
- background subtraction
- image denoising
- d scene
- map estimation
- image pairs
- low resolution
- multiresolution
- image analysis
- moving objects
- three dimensional