Motivation

Objectives

• We propose a real-time and lightweight single-image super-resolution (SR) network named Bicubic++. 

  1. Despite using spatial dimensions of the input image across the whole network, Bicubic++ first learns quick reversible downgraded and lower resolution features of the image in order to decrease the number of computations. 

  2. We also construct a training pipeline, where we apply an end-to-end global structured pruning of convolutional layers without using metrics like magnitude and gradient norms, and focus on optimizing the pruned network's PSNR on the validation set. 

• 
  
• 
  

Related Works

Idea

Steps

Architecture

Data Augmentation

Encoder

Decoder

Heads

Loss Function

Training Strategy

Dataset

Metrics

Experimental Results

Ablations