peylnog / crdnet Goto Github PK

27.0 0.0 8.0 63 KB

Cumulative Rain Density Sensing Network for Single Image Derain(CRDNet) IEEE SPL 2020

Home Page: https://github.com/peylnog/CRDNet

Python 100.00%

crdnet's Introduction

Cumulative Rain Density Sensing Network for Single Image Derain

Abstract

This paper focuses on single image derain, which aims to restore clear image from single rain image. Through full consideration of different frequency information preservation and the complicated interactions between rain-streaks and background, a novel end-to-end cumulative rain-density sensing network (CRDNet) is proposed for adaptive rain-streaks removal. An effective W-Net with powerful learning ability is proposed as a key component to recover rain-invariant low-frequency signals. A cumulative rain-density classifier with a novel cost-sensitive label encoding strategy is proposed as an auxiliary network to improve discriminative power of extracted high-frequency rain-streaks through multi-task training. The proposed CRDNet has been compared with state-of-the-art methods on two public datasets. The quantitative and visual experimental results demonstrate that it can achieve excellent performance with great improvement. Related source code and models are available on github https://github.com/peylnog/CRDNet.

✨ Getting Start

Installation

Framework

Python 3.7
Pytorch1.0 (with ubuntu 16)
Torchvision

Python Dependencies

skimage
numpy
visdom : pip install visdom

Train

Download Weights BaiDuYunLink passwd:ncf3

python3.7 derain.py

Test

python3.7 derain_test.py
ps:make sure data root is right

Citation

Please cite us if this work is helpful to you.

    InProceedings{
    author = {L. Peng, A. Jiang, Q. Yi and M. Wang},
    title = {Cumulative Rain Density Sensing Network for Single Image Derain(CRDNet)},
    booktitle = {IEEE Signal Processing Letters(SPL)},
    month = {February},
    year = {2020}
    }

crdnet's People

Contributors

Stargazers

Forkers

joyies cv-ip northicesheep tkone2018 perseverancelx madeline-hyh gatsby-bot aw-jiang

crdnet's Issues

Concerning the Inauthentic Increase in Stars for CRDNet Repository

Dear Maintainers,

I hope this message finds you well. I am writing to express my deep concern regarding the recent and substantial increase in the number of stars for the CRDNet repository. Upon close inspection, it is evident that this surge is highly irregular and appears to be the result of artificial manipulation rather than genuine community interest.

As a seasoned researcher and long-time GitHub user, it is disheartening to witness such practices. The integrity of our community relies heavily on the authenticity of contributions and the transparent sharing of knowledge. Engaging in or endorsing the inflation of metrics through inauthentic means undermines the trust and collaborative spirit that GitHub is built upon.

Here are a few points that substantiate my concerns:

Sudden Surge After Years of Stability: The CRDNet repository has been available for over five years with a relatively stable and modest number of stars. The abrupt spike in popularity, without any significant updates or breakthroughs, raises questions about the legitimacy of this growth.
Quality and Impact: While the project is competent, it does not exhibit groundbreaking innovation or substantial improvements that would naturally attract such a dramatic increase in attention. The quality and impact of a repository typically correlate with its popularity, and in this case, the correlation appears to be artificially skewed.
Community Trust: Artificially boosting repository metrics erodes trust within the GitHub community. It creates an unfair playing field where projects with genuine merit may be overshadowed by those employing deceptive tactics. This behavior is particularly frustrating for researchers who rely on authentic feedback and engagement to gauge the relevance and impact of their work.

I have already reported this issue to the GitHub support team and the broader community for further investigation. It is crucial that we maintain a fair and transparent environment to ensure that the platform remains a reliable resource for developers, researchers, and contributors worldwide.

I urge you to address this issue promptly and take appropriate measures to rectify the situation. Transparency and honesty are paramount in our field, and I hope you will align with these values moving forward.

Thank you for your attention to this matter!

ValueError: num_samples should be a positive integer value, but got num_samples=0

When i run your code,it shows ValueError,as show in the follow image

Besides ,could you show me your data root and path code in your DefaultConfig() as well as some code in your test(). The three parts as my follow images.

Bug：name 'conv_blocks_size_5' is not defined

Your paper ‘Cumulative Rain Density Sensing Network for Single Image Derain ’ is in line with my research direction. There are many places in the paper gave me great reference value. However, when I tested your code, I found some errors in your code。
Hope you can check the code and look forward to your reply