GitPlanet

Cheap and reliable Node.js hosting starts at $3/month, and $1/month static HTML hosting

Created with love in Canada, visit hostnodejs.com today

Feel like to post an Ad? Learn Details
All Projects → lizhengwei1992 → IDN-pytorch

lizhengwei1992 / IDN-pytorch

Licence: other
paper implement : Fast and Accurate Single Image Super-Resolution via Information Distillation Network

Programming Languages

python
139335 projects - #7 most used programming language

Labels

pytorchimage-super-resolutioncvpr2018

Projects that are alternatives of or similar to IDN-pytorch

Cvpr18 Inaturalist Transfer
Large Scale Fine-Grained Categorization and Domain-Specific Transfer Learning. CVPR 2018
Stars: ✭ 164 (+310%)
Mutual labels:  cvpr2018
NLRN
Code for Non-Local Recurrent Network for Image Restoration (NeurIPS 2018)
Stars: ✭ 165 (+312.5%)
Mutual labels:  image-super-resolution
Awesome-low-level-vision-resources
A curated list of resources for Low-level Vision Tasks
Stars: ✭ 35 (-12.5%)
Mutual labels:  image-super-resolution
Sin
CVPR 2018: Structure Inference Net for Object Detection
Stars: ✭ 178 (+345%)
Mutual labels:  cvpr2018
Image manipulation detection
Paper: CVPR2018, Learning Rich Features for Image Manipulation Detection
Stars: ✭ 210 (+425%)
Mutual labels:  cvpr2018
FSRCNN-pytorch
PyTorch implementation of Accelerating the Super-Resolution Convolutional Neural Network (ECCV 2016)
Stars: ✭ 115 (+187.5%)
Mutual labels:  image-super-resolution
Frvsr
Frame-Recurrent Video Super-Resolution (official repository)
Stars: ✭ 157 (+292.5%)
Mutual labels:  cvpr2018
FaceAttr
CVPR2018 Face Super-resolution with supplementary Attributes
Stars: ✭ 18 (-55%)
Mutual labels:  cvpr2018
Dfl Cnn
This is a pytorch re-implementation of Learning a Discriminative Filter Bank Within a CNN for Fine-Grained Recognition
Stars: ✭ 245 (+512.5%)
Mutual labels:  cvpr2018
Image Super Resolution
🔎 Super-scale your images and run experiments with Residual Dense and Adversarial Networks.
Stars: ✭ 3,293 (+8132.5%)
Mutual labels:  image-super-resolution
Attentive Gan Derainnet
Unofficial tensorflow implemention of "Attentive Generative Adversarial Network for Raindrop Removal from A Single Image (CVPR 2018) " model https://maybeshewill-cv.github.io/attentive-gan-derainnet/
Stars: ✭ 184 (+360%)
Mutual labels:  cvpr2018
Guided Attention Inference Network
Contains implementation of Guided Attention Inference Network (GAIN) presented in Tell Me Where to Look(CVPR 2018). This repository aims to apply GAIN on fcn8 architecture used for segmentation.
Stars: ✭ 204 (+410%)
Mutual labels:  cvpr2018
SRCNN-pytorch
PyTorch implementation of Image Super-Resolution Using Deep Convolutional Networks (ECCV 2014)
Stars: ✭ 176 (+340%)
Mutual labels:  image-super-resolution
Dbnet
DBNet: A Large-Scale Dataset for Driving Behavior Learning, CVPR 2018
Stars: ✭ 172 (+330%)
Mutual labels:  cvpr2018
glimpse clouds
Pytorch implementation of the paper "Glimpse Clouds: Human Activity Recognition from Unstructured Feature Points", F. Baradel, C. Wolf, J. Mille , G.W. Taylor, CVPR 2018
Stars: ✭ 30 (-25%)
Mutual labels:  cvpr2018
L Gm Loss
Implementation of our accepted CVPR 2018 paper "Rethinking Feature Distribution for Loss Functions in Image Classification"
Stars: ✭ 162 (+305%)
Mutual labels:  cvpr2018
HCFlow
Official PyTorch code for Hierarchical Conditional Flow: A Unified Framework for Image Super-Resolution and Image Rescaling (HCFlow, ICCV2021)
Stars: ✭ 140 (+250%)
Mutual labels:  image-super-resolution
RDN-pytorch
PyTorch implementation of Residual Dense Network for Image Super-Resolution (CVPR 2018)
Stars: ✭ 91 (+127.5%)
Mutual labels:  image-super-resolution
PESR
Official code (Pytorch) for paper Perception-Enhanced Single Image Super-Resolution via Relativistic Generative Networks
Stars: ✭ 28 (-30%)
Mutual labels:  image-super-resolution
SwinIR
SwinIR: Image Restoration Using Swin Transformer (official repository)
Stars: ✭ 1,260 (+3050%)
Mutual labels:  image-super-resolution
View All Similar Projects ➔

IDN-pytorch

Pytorch implement of the paper, Fast and Accurate Single Image Super-Resolution via Information Distillation Network. You can get the caffe code and pretrained model given by author.

I keep the architecture network in pytorch same as the paper described, ignoring some details.

Analysis

The proposed network learns residual to reconstruct HR image. architecture of network Imformation distillation means that the proposed enhancement unit mixes together two different types of features and the compression unit distills more useful information for the sequential blocks.
Features sliced from middle layer concatenate with features previous layer, and the result add to subsequent layer, this likes another residual dense connect in my opion.

architecture of enhancement unitcompress

The compression layer just reduce dimensions of channels using 1x1 convolution, same as dense net.

As I know, many proposed network want to reuse previous layers as more as possiable, inclouding this paper. The RDN(paper and code) network seems more efficent.

Requirements

  • python3.5 / 3.6
  • pytorch >= 0.2
  • opencv

Train

python3 main.py --model_name IDN --load demo_IDN --dataDir ./data/ --need_patch True --blur True --patchSize 288 --nFeat_slice 4 --nFeat 64 --nDiff 16 --scale 3 --epoch 5000 --lrDecay 2000 --lr 1e-4 --batchSize 16 --lossType 'L1' --nThreads 6
Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].