/SPMC_VideoSR-pytorch

PyTorch Implementation of [Detail-revealing Deep Video Super-resolution](https://arxiv.org/abs/1704.02738)

Primary LanguagePython

Overview

TODOs

  • Tensorboard Support
  • GIF
SPMC_VideoSR
├── cfgs
│   ├── config_latest.py
│   └── config.py
├── data
│   └── test
│       └── 1.jpg
├── dataset.py
├── losses.py
├── model.py
├── modules
│   ├── detail_fusion_net.py
│   ├── __init__.py
│   ├── me.py
│   └── spmc.py
├── np.py
├── predict.py
├── preprocess.py
├── utils
│   ├── flow.py
│   ├── color.py
│   ├──
└── WORKBENCH.ipynb

Usage

  • Preprocess
  • Train
  • Predict
  • Load model

Results

Image Results

Input Output Groundtruth

Metrics

(PSNR/SSIM)

Method Ours Paper
×2 36.71 / 0.96
×3 31.92 / 0.90
×4 29.69 / 0.84

  • Our flow visualization is different from MPI-Sintel's. for more details about flow estimation.
  • Use ffmpeg to make GIFs

References