fu123456 / shiq Goto Github PK
View Code? Open in Web Editor NEWProject for CVPR21 paper: "A Multi-Task Network for Joint Specular Highlight Detection and Removal".
Project for CVPR21 paper: "A Multi-Task Network for Joint Specular Highlight Detection and Removal".
Hello,
Could you provide the source code for your model ?
The .so is useful for testing as-is but I'd like to mess with the parameters myself.
Thanks in advance
您好!请问如何得到A对应的D,S,M图像,以制作自己的数据集呀?
请问下,我的环境配置完成,但是提示RuntimeError,我是在虚拟机上进行运行。
Originally posted by @JacobYeah in #3 (comment)
how to run the matlab code? where is the dataface metioned by load('./data/dataFace.mat') in sidDemo
We noticed that you posted a test model. Before deciding whether to go further, the researcher seems to want to see the test image directly. Can you provide test results? For example, a zip file containing test images (.png
) available for download. Looking forward to your reply.
in JSHDR paper, you use RPCA to generate the highlight-free image from highlight image as the removal ground truth, if the RPCA can remove highlight well,why we need a new mothod of JSHDR?
such as?
调用这个包没法加载
to 大神
作者的工作非常棒~
我想将它应用到下游任务中,但是由于原始图像本身分辨率太高,直接应用权重的移除效果并不好,但是它检测的结果很不错。
不知道作者什么时候可以提供一个MASK输出的接口。非常感谢。
请问这个错误应该怎么修改呀
你好同学,最近在复现您的这篇工作,有一些细节的地方图不是很清楚想请教一下
一共有两个问题
第一个问题,输入图片到第一个DSCFA时,中间是否要加入一些cnns 来改变通道数(3 ->k)
第二个问题,F经过convs得到M时,这里的M是指的一个特征图还是一个展示出来的detect_mask?(S同理?)因为mask的通道应该是1吧,而M后续还有两步的使用,感觉应该M是一个k通道的特征图而不是最后的mask,mask应该是在M的基础上再加上了一个卷积得到的?
第三个问题, 如果第二问题M和S都是k个通道的特征图的话,那I-MS计算的时候,I是不是要先经过卷积通道也做到k?
im a junior, wanna know how to get detectRemoval document from the detectionRemoval.cpython-36m-x86_64-linux-gnu.so ?
按照要求安装环境,我的显卡cuda是11.6,但是安装pytorch1.5.0只有cuda10.2的版本,运行代码后出现如下错误:RuntimeError: CUDA error: no kernel image is available for execution on the device。
用其它版本的pytorch跑又会出现其它的错误。
请问作者怎么解决?还有是否可以提供源码?
希望得到您的回复
Hello,
I'm trying to reimplement your network but having problem on the DSCFA module. In the paper you wrote "Our DSCFA block learns dilated spatial features to extract and aggregate dilated contextual features from four directions" and the figure showed in Figure 7.
I'm wondering:
Thanks!
Can you tell me how to use LIME dataset to train and test a model?
On the one hand, which one is the highlighted folder, and which one is the unhighlighted folder.
On the other hand, LIME only has training data, which pictures to use for testing.
Looking forward to your reply! Thanks!
您好,我下载模型后用infer.py代码测试,
代码内:
from utils.detectionRemoval import test_single_highlight_image
工程内未找到detectionRemoval 文件,只有detectionRemoval.cpython-36m-x86_64-linux-gnu.so文件,
是工程缺少文件,还是本人使用错误。
感谢指导
请问,无GPU的电脑能跑吗?需要修改哪处代码?
可以开源训练代码吗?我如果想训练自己的数据集,可以吗?谢谢
你好,我想下载您的训练模型进行测试,
但是谷歌网盘下载需要权限。
能麻烦您给一个能下载的模型吗?
感谢
您好,请问可以告诉一下参数量吗?谢谢您。
您好。我想制作自己的数据集,目前只有两种类型的图片(_A.png 和 _D.png),您提供的开源代码中有生成_T.png的工具,但是需要用到_S.png,那么请问这个_S.png是怎样获得的 ?我在您的文章里没有看到有提到(或者是我没有理解)。
您好,在论文中您提到的在SHIQ数据集上的去高光结果是PSNR=34.13,SSIM=0.86,请问计算的时候是只在的高光区域上进行计算,还是在整张图片上计算的?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.