Model Framework

Pre-upsampling SR

Post-upsampling SR

Progressive Upsampling SR

Iterative Up-and-down upsampling SR

Upsampling Methods

Interpolation-based methods

Nearest Neighbor

Bilinear

Bicubic

Others

Learning-based Methods

Transposed Convolution

Sub-pixel Layer

Meta Upscale Module

Network Design

Residual Learning

Recursive Learning

Multi-path Learning

Dense Connection

Attention Mechanism

Advanced Convolution

Region-recursive Learning

Pyramid Pooling

Wavelet Transformation

xUnit

Desubpixel

Learning Strategies

Loss Functions:

-  Pixel Loss

-  Content Loss

-  Texture Loss

-  Adversarial Loss

-  Cycle Consistency Loss

-  Total Variation Loss

-  Prior-based Loss

Batch Normalization

Curriculum Learning

Multi-supervision

Other Improvement

Context-wise Network Fusion

Data Augmentation

Multi-task Learning

Network Interpolation

Self-ensemble