Weiwei's starred repositories
Real-ESRGAN
Real-ESRGAN aims at developing Practical Algorithms for General Image/Video Restoration.
vit-pytorch
Implementation of Vision Transformer, a simple way to achieve SOTA in vision classification with only a single transformer encoder, in Pytorch
image-super-resolution
🔎 Super-scale your images and run experiments with Residual Dense and Adversarial Networks.
Image-Super-Resolution-via-Iterative-Refinement
Unofficial implementation of Image Super-Resolution via Iterative Refinement by Pytorch
Palette-Image-to-Image-Diffusion-Models
Unofficial implementation of Palette: Image-to-Image Diffusion Models by Pytorch
Image-Adaptive-3DLUT
Learning Image-adaptive 3D Lookup Tables for High Performance Photo Enhancement in Real-time
ddpm-segmentation
Label-Efficient Semantic Segmentation with Diffusion Models (ICLR'2022)
Harmonizer
High-Resolution Image/Video Harmonization [ECCV 2022]
Style-Your-Hair
Official Pytorch implementation of "Style Your Hair: Latent Optimization for Pose-Invariant Hairstyle Transfer via Local-Style-Aware Hair Alignment (ECCV 2022)"
2022-CVPR-AirNet
PyTorch implementation for All-In-One Image Restoration for Unknown Corruption (AirNet) (CVPR 2022)
Chinese-Character-and-Calligraphic-Image-Processing
Some interesting method like style transfer, GAN, deep neural networks for Chinese character and calligraphic image processing
Video-Harmonization-Dataset-HYouTube
[IJCAI 2022] The first public benchmark dataset for video harmonization. The code used in our paper "Deep Video Harmonization with Color Mapping Consistency", IJCAI 2022.
DeepSteganography
Ever sent a hidden message in invisible ink to your friends? Are you intrigued by the idea of cryptic message exchange? How about using images for this exchange? Steganography is what you need! It is one of the techniques of encryption and over the years, steganography has been used to encode a lower resolution image into a higher resolution image. But steganography using naive methods, like LSB manipulation, is susceptible to statistical analysis. Our model extends existing deep learning research for encoding multiple secret images onto a single cover by leveraging convolutional neural networks based deep learning architectures. DeepSteg allows senders to embed up to three secret images onto a single cover using an encoder network and then have multiple decoder networks to obtain the embedded secrets.
DeshadowSTCGANs
An implementation of the paper:"Stacked Conditional Generative Adversarial Networks for Jointly Learning Shadow Detection and Shadow Removal" by Jifeng Wang, etc.