DeT

We propose a transformer-based method for deinterlacing. The method consists of three main modules, the Feature extractor, Deinterlacing transformer and Densnet modules.

Installation

• Python >= 3.7 (Recommend to use Anaconda or Miniconda)

• PyTorch >= 1.3

• NVIDIA GPU + CUDA Clone repository

  git clone https://github.com/PGAnonymous2022/DeT.git

Dataset Preparation

• Please refer to paper for more details.

Training

• Please refer to main.py

  # Train on Vimeo-90K
  CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 python main.py

Testing

• Please refer to configuration of testing for more details.

  # Test on Vimeo-90K
  CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 python demo_Vid4.py