Unsupervised Image-to-Image Translation Networks --- Reading Writing

2017.03.03

  Motivations: most existing image to image translation algorithms are all need image pairs as training data for deep neural network, such as CGANs or VAEs. But in some cases, it is rather difficult to collect such training data. For example, the night and day image pairs, the perfect aligned thermal RGB image pairs, or sunning rainning, fogging, et al, which provide us a new challenging problem:

  How to do image to image translation in a unsupervised fashion which do not need aligned image pairs ?

This paper proposed the UNIT framework (UNsupervised Image-to-image Translation network) to deal with this problem which combine VAE and GANs. The whole framework can be described as the following figures which seems complex but rather easy to understand.

  

  There are two most important assumptions about the proposed framework:

  1. we assume that the relationship between X1 and X2 does not only exist at the image level but also at the level of local patches or regions.

  2. for any given images x1 and x2,  there exists a common underlying representation z, such that we can cover both images from this underlying representation from each of the two input images.

  VAEs: the encoder-generator pair {E1, G1} constitutes a VAE for the X1 domain, termed VAE1. Another pair of {E2, G2} constitutes a VAE for the X2 domain VAE2.

    Weight-sharing : we enforce a weight-sharing constraint to relate the representations in the two VAEs.

  GANs :  two GANs are used to output the two domains.

  

  

  

  


    Experiments:

  

  

    

  

  

  

Unsupervised Image-to-Image Translation Networks --- Reading Writing的更多相关文章

  1. Learning to Compare Image Patches via Convolutional Neural Networks --- Reading Summary

    Learning to Compare Image Patches via Convolutional Neural Networks ---  Reading Summary 2017.03.08 ...

  2. Unsupervised Image-to-Image Translation Networks

    Abstract: 无监督图像到图像的翻译目的是学习不同域图像的一个联合分布,通过使用来自单独域图像的边缘分布.给定一个边缘分布,可以得到很多种联合分布.如果不加入额外的假设条件的话,从边缘分布无法推 ...

  3. reading/writing files in Python

    file types: plaintext files, such as .txt .py Binary files, such as .docx, .pdf, iamges, spreadsheet ...

  4. 【Deep Learning】Hinton. Reducing the Dimensionality of Data with Neural Networks Reading Note

    2006年,机器学习泰斗.多伦多大学计算机系教授Geoffery Hinton在Science发表文章,提出基于深度信念网络(Deep Belief Networks, DBN)可使用非监督的逐层贪心 ...

  5. On Explainability of Deep Neural Networks

    On Explainability of Deep Neural Networks « Learning F# Functional Data Structures and Algorithms is ...

  6. The Unreasonable Effectiveness of Recurrent Neural Networks (RNN)

    http://karpathy.github.io/2015/05/21/rnn-effectiveness/ There’s something magical about Recurrent Ne ...

  7. 26 THINGS I LEARNED IN THE DEEP LEARNING SUMMER SCHOOL

    26 THINGS I LEARNED IN THE DEEP LEARNING SUMMER SCHOOL In the beginning of August I got the chance t ...

  8. DotNet 资源大全中文版(Awesome最新版)

    Awesome系列的.Net资源整理.awesome-dotnet是由quozd发起和维护.内容包括:编译器.压缩.应用框架.应用模板.加密.数据库.反编译.IDE.日志.风格指南等. 算法与数据结构 ...

  9. (转) Awesome - Most Cited Deep Learning Papers

    转自:https://github.com/terryum/awesome-deep-learning-papers Awesome - Most Cited Deep Learning Papers ...

随机推荐

  1. sitecore系统教程之使用修补程序文件自定义Sitecore配置

    您可以使用修补程序文件在Sitecore中添加或更改配置设置.Sitecore将修补程序文件与Sitecore.config 文件合并,以创建在运行时使用的配置文件. 本主题描述: 补丁文件放置 补丁 ...

  2. 【impala学习之一】impala

    环境 虚拟机:VMware 10 Linux版本:CentOS-6.5-x86_64 客户端:Xshell4 FTP:Xftp4 jdk8 CM5.4 一.ImpalaImpala是基于Hive的大数 ...

  3. linux 查看python安装路径,版本号

    一.想要查看ubuntu中安装的python路径 方法一:whereis python     方法二:which python   二.想要查看ubuntu中安装的python版本号 python ...

  4. Log4j基础知识

    Log4J是Apache的一个开放源代码的项目.通过使用Log4J,程序员可以控制日志信息输送的目的地,包括控制台,文件,GUI组件和NT事件记录器,也可以控制每一条日志的输出格式,或通过定义每一条日 ...

  5. windows.onload和body的onload属性的区别

    关于windows.onload和body的onload属性的区别网上有些说法说的也不太统一,现在系统说下: 先看共同点: 都是body内容体加载结束执行: window.onload 内部方式可以 ...

  6. SharePoint入门——创建一个网站

    1.首先安装SP相关环境,可以百度到具体操作步骤: (以下步骤基于本人环境:本人用的Win10自带的Hyper-V虚拟机.Windows Server2012R2.SQL2014以及SharePoin ...

  7. Python智能检测编码并转码

    #安装包工具 $pip3 install chardet #直接打开文件,中文显示乱码 >>> import chardet >>> f = open('test. ...

  8. Django框架----Object Relational Mapping(ORM)

    Django中的ORM Django项目使用MySQL数据库 1. 在Django项目的settings.py文件中,配置数据库连接信息: DATABASES = { "default&qu ...

  9. 每日linux命令学习-grep模式检索

    grep模式检索指令包括grep,egrep,和fgrep,.Linux系统使用正则表达式优化文本检索,所以在此,笔者首先学习了一下正则表达式. 1. 正则表达式 正则表达式使用被称为元字符(Meta ...

  10. 登录实现之servlet和SpringMVC

    #知识小点: 防止乱码的字符集的设置: request.setCharacterEncoding():确定从请求端发送到给服务器的编码是汉字字符集.该方法对get方法 无效,只对post方法有效.若要 ...