BiSTNet: Semantic Image Prior Guided Bidirectional Temporal Feature Fusion for Deep Exemplar-based Video Colorization
This repository is the official pytorch implementation of our paper, BiSTNet: Semantic Image Prior Guided Bidirectional Temporal Feature Fusion for Deep Exemplar-based Video Colorization.
Yixin Yang1,
Zhongzheng Peng1,
Xiaoyu Du1,
Zhulin Tao2,
Jinhui Tang1,
Jinshan Pan1
1Nanjing University of Science and Technology, 2Communication University of China
We are excited to announce that our method based on BiSTNet won 1st place in the NTIRE2023 Video Colorization Challenge Track 1, Fréchet Inception Distance (FID) Optimization! You can find our code and the pre-trained weights in the following links:
- Release training code
- Release testing code
- Release pre-trained models
This repo is currently maintained by Yixin Yang (@yyang181) and is for academic research use only.