Come-Closer-Diffuse-Faster: Accelerating Conditional Diffusion Models for Inverse Problems through Stochastic Contraction.
Hyungjin ChungByeongsu SimJong Chul YePublished in: CVPR (2022)
Keyphrases
- inverse problems
- diffusion models
- image reconstruction
- global optimization
- diffusion model
- convex optimization
- information diffusion
- optimization methods
- social networks
- optimization problems
- early vision
- influence maximization
- partial differential equations
- greedy algorithm
- computationally expensive
- high quality
- viral marketing
- anisotropic diffusion
- diffusion process
- particle swarm optimization
- smoothness constraint
- medical images
- image sequences
- image processing
- computer vision