The implementation of ICASSP 2020 paper "Pixel-level self-paced learning for super-resolution"

Elon Lin

Last update: Dec 15, 2022

Related tags

Overview

Pixel-level Self-Paced Learning for Super-Resolution

This is an official implementaion of the paper Pixel-level Self-Paced Learning for Super-Resolution, which has been accepted by ICASSP 2020.

[arxiv][PDF]

trained model files: Baidu Pan(code: v0be)

Requirements

This code is forked from thstkdgus35/EDSR-PyTorch. In the light of its README, following libraries are required:

Python 3.6+ (Python 3.7.0 in my experiments)
PyTorch >= 1.0.0 (1.1.0 in my experiments)
numpy
skimage
imageio
matplotlib
tqdm

Core Parts

Detail code can be found in Loss.forward, which can be simplified as:

# take L1 Loss as example

import torch
import torch.nn as nn
import torch.nn.functional as F
from . import pytorch_ssim

class Loss(nn.modules.loss._Loss):
    def __init__(self, spl_alpha, spl_beta, spl_maxVal):
        super(Loss, self).__init__()
        self.loss = nn.L1Loss()
        self.alpha = spl_alpha
        self.beta = spl_beta
        self.maxVal = spl_maxVal

    def forward(self, sr, hr, step):
        # calc sigma value
        sigma = self.alpha * step + self.beta
        # define gauss function
        gauss = lambda x: torch.exp(-((x+1) / sigma) ** 2) * self.maxVal
        # ssim value
        ssim = pytorch_ssim.ssim(hr, sr, reduction='none').detach()
        # calc attention weight
        weight = gauss(ssim).detach()
        nsr, nhr = sr * weight, hr * weight
        # calc loss
        lossval = self.loss(nsr, nhr)
        return lossval

the library pytorch_ssim is focked from Po-Hsun-Su/pytorch-ssim and rewrite some details for adopting it to our requirements.

Attention weight values change according to SSIM Index and steps:

Citation

If you find this project useful for your research, please cite:

@inproceedings{lin2020pixel,
  title={Pixel-Level Self-Paced Learning For Super-Resolution}
  author={Lin, Wei and Gao, Junyu and Wang, Qi and Li, Xuelong},
  booktitle={ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  year={2020},
  pages={2538-2542}
}

Pytorch implementation for the EMNLP 2020 (Findings) paper: Connecting the Dots: A Knowledgeable Path Generator for Commonsense Question Answering

Path-Generator-QA This is a Pytorch implementation for the EMNLP 2020 (Findings) paper: Connecting the Dots: A Knowledgeable Path Generator for Common

33 Dec 5, 2022

PyTorch implementation of ECCV 2020 paper "Foley Music: Learning to Generate Music from Videos "

Foley Music: Learning to Generate Music from Videos This repo holds the code for the framework presented on ECCV 2020. Foley Music: Learning to Genera

30 Nov 3, 2022

The pytorch implementation of the paper "text-guided neural image inpainting" at MM'2020

TDANet: Text-Guided Neural Image Inpainting, MM'2020 (Oral) MM | ArXiv This repository implements the paper "Text-Guided Neural Image Inpainting" by L

75 Dec 22, 2022

PyTorch implementation of CVPR 2020 paper (Reference-Based Sketch Image Colorization using Augmented-Self Reference and Dense Semantic Correspondence) and pre-trained model on ImageNet dataset

Reference-Based-Sketch-Image-Colorization-ImageNet This is a PyTorch implementation of CVPR 2020 paper (Reference-Based Sketch Image Colorization usin

11 Jul 28, 2022

Code for ACM MM 2020 paper "NOH-NMS: Improving Pedestrian Detection by Nearby Objects Hallucination"

NOH-NMS: Improving Pedestrian Detection by Nearby Objects Hallucination The offical implementation for the "NOH-NMS: Improving Pedestrian Detection by

64 Nov 11, 2022

TensorFlow code for the neural network presented in the paper: "Structural Language Models of Code" (ICML'2020)

SLM: Structural Language Models of Code This is an official implementation of the model described in: "Structural Language Models of Code" [PDF] To ap

73 Nov 6, 2022

Code for our paper at ECCV 2020: Post-Training Piecewise Linear Quantization for Deep Neural Networks

PWLQ Updates 2020/07/16 - We are working on getting permission from our institution to release our source code. We will release it once we are granted

54 Dec 15, 2022

A code repository associated with the paper A Benchmark for Rough Sketch Cleanup by Chuan Yan, David Vanderhaeghe, and Yotam Gingold from SIGGRAPH Asia 2020.

A Benchmark for Rough Sketch Cleanup This is the code repository associated with the paper A Benchmark for Rough Sketch Cleanup by Chuan Yan, David Va

33 Dec 18, 2022

Code for the paper: Adversarial Training Against Location-Optimized Adversarial Patches. ECCV-W 2020.

Adversarial Training Against Location-Optimized Adversarial Patches arXiv | Paper | Code | Video | Slides Code for the paper: Sukrut Rao, David Stutz,

32 Dec 13, 2022

Comments

Could you please specify the usage of your code? THX

Hi Elin

Thanks for your work, recently I want to reproduce the result of your method. Can you specify the work-flow of your paper, for convenient usage for others to use.

Thanks !

opened by JustinAsdz 4

Question regarding the effect of the similarity map

I was curious as to what the effect of the similarity map was, so I added a few lines of code to the forward function of the Loss class, to write out sr[0] and hr[0] patches before and after multiplication by weight=gauss(ssim).detach(), for batch=1 of each epoch. My training command was:

python main.py --model EDSR --scale 4 --data_test Set5+Set14+B100+Urban100+DIV2K --n_GPUs 1 --epochs 300

For clarification, all arguments are:

Namespace(G0=64, RDNconfig='B', RDNkSize=3, act='relu', batch_size=16, betas=(0.9, 0.999), chop=False, cpu=False, data_range='1-800/801-900', data_test=['Set5', 'Set14', 'B100', 'Urban100', 'DIV2K'], data_train=['DIV2K'], debug=False, decay='200', dilation=False, dir_data='../x_imagedata', dir_demo='../test', disable_PSPL=False, epochs=300, epsilon=1e-08, ext='sep', extend='.', gamma=0.5, gan_k=1, gclip=0, load='', loss='1*L1', lr=0.0001, model='EDSR', momentum=0.9, n_GPUs=1, n_colors=3, n_feats=64, n_layers=8, n_resblocks=16, n_resgroups=10, n_threads=6, negative_slope=0.2, no_augment=False, optimizer='ADAM', patch_size=192, pre_train='', precision='single', print_every=250, reduction=16, res_scale=1, reset=False, resume=0, rgb_range=255, save='EDSR_04-08_22-15-40', save_gt=False, save_models=False, save_results=False, scale=[4], seed=1, self_ensemble=False, shift_mean=True, skip_threshold=100000000.0, splalpha=0.3, splbeta=0, split_batch=1, splval=2, template='.', test_every=1000, test_only=False, weight_decay=0)

My evaluation results were;

  [Set5 x4]     PSNR: 32.076 (Best: 32.134 @epoch 268)  ssim=0.896102
  [Set14 x4]    PSNR: 28.535 (Best: 28.568 @epoch 267)  ssim=0.785463
  [B100 x4]     PSNR: 27.539 (Best: 27.547 @epoch 257)  ssim=0.743243
  [Urban100 x4] PSNR: 25.956 (Best: 25.961 @epoch 293)  ssim=0.785183
  [DIV2K x4]    PSNR: 28.897 (Best: 28.903 @epoch 257)  ssim=0.837567

Here is what the images look like, as the epochs change, for just a few epochs. From left to right, these are sr[0], hr[0], sr[0]*weight, hr[0]*weight.

Is this about what you'd expect? They just seemed a little noisier to me than, e.g., Figure 2 in the paper. I can also try the training commands that you used in #1; the x2 case is running now, it looks like it'll take about 2.5 days...

opened by drcdr 1

The implementation of ICASSP 2020 paper "Pixel-level self-paced learning for super-resolution"

Related tags

Overview

Pixel-level Self-Paced Learning for Super-Resolution

Requirements

Core Parts

Citation

You might also like...

Pytorch implementation for the EMNLP 2020 (Findings) paper: Connecting the Dots: A Knowledgeable Path Generator for Commonsense Question Answering

PyTorch implementation of ECCV 2020 paper "Foley Music: Learning to Generate Music from Videos "

The pytorch implementation of the paper "text-guided neural image inpainting" at MM'2020

PyTorch implementation of CVPR 2020 paper (Reference-Based Sketch Image Colorization using Augmented-Self Reference and Dense Semantic Correspondence) and pre-trained model on ImageNet dataset

Code for ACM MM 2020 paper "NOH-NMS: Improving Pedestrian Detection by Nearby Objects Hallucination"

TensorFlow code for the neural network presented in the paper: "Structural Language Models of Code" (ICML'2020)

Code for our paper at ECCV 2020: Post-Training Piecewise Linear Quantization for Deep Neural Networks

A code repository associated with the paper A Benchmark for Rough Sketch Cleanup by Chuan Yan, David Vanderhaeghe, and Yotam Gingold from SIGGRAPH Asia 2020.

Code for the paper: Adversarial Training Against Location-Optimized Adversarial Patches. ECCV-W 2020.

Comments

Could you please specify the usage of your code? THX

Question regarding the effect of the similarity map

Owner

Elon Lin

Code for our ICASSP 2021 paper: SA-Net: Shuffle Attention for Deep Convolutional Neural Networks

Code for the paper "Unsupervised Contrastive Learning of Sound Event Representations", ICASSP 2021.

Code for the ICASSP-2021 paper: Continuous Speech Separation with Conformer.

Official implementation of FCL-taco2: Fast, Controllable and Lightweight version of Tacotron2 @ ICASSP 2021

This is the implementation of "SELF SUPERVISED REPRESENTATION LEARNING WITH DEEP CLUSTERING FOR ACOUSTIC UNIT DISCOVERY FROM RAW SPEECH" submitted to ICASSP 2022

UDP++ (ECCVW 2020 Oral), (Winner of COCO 2020 Keypoint Challenge).

Code for paper ECCV 2020 paper: Who Left the Dogs Out? 3D Animal Reconstruction with Expectation Maximization in the Loop.

PyTorch implementation of the Deep SLDA method from our CVPRW-2020 paper "Lifelong Machine Learning with Deep Streaming Linear Discriminant Analysis"

A PyTorch Implementation of the paper - Choi, Woosung, et al. "Investigating u-nets with various intermediate blocks for spectrogram-based singing voice separation." 21th International Society for Music Information Retrieval Conference, ISMIR. 2020.

The implementation of the algorithm in the paper "Safe Deep Semi-Supervised Learning for Unseen-Class Unlabeled Data" published in ICML 2020.