TTSR_b2

An Improved Texture Transformer Network for Ref-based Image Super-Resolution.

1.Introduciton

This is a project to improve the TTSR method. We nonlinearly enhance the attention mechanism. The official PyTorch implementation of the paper Learning Texture Transformer Network for Image Super-Resolution is in here.

LR Image

SR Image

Ref Image

We compare our method TTSRA with TTSR, and the experimental results show that our method performs better. Our method has higher recovery ability for image details.

2.Method

3.Preparation

(1). Download CUFED train set and CUFED test set (2). Make dataset structure be:

• CUFED
  • train
    • input
    • ref
  • test
    • CUFED5

4.Train & Evaluation

python3 main.py

Please use numbers in annotation to choose the function you need.