<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[yl]]></title><description><![CDATA[自动驾驶]]></description><link>https://blog.youkuaiyun.com/xumi13</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; xumi13]]></copyright><item><title><![CDATA[LineNet - a Zoomable CNN for Crowdsourced High Definition Maps Modeling in Urban Environments]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/105443876</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/105443876</guid><author>xumi13</author><pubDate>Fri, 10 Apr 2020 22:38:33 +0800</pubDate><description><![CDATA[一.概述
时间：2018.07
作者：Dun Liang
机构：Tsinghua University
内容：LineNet， a convolutional neural network with a novel prediction layer and a zoom module
数据集：TTLane
二.方法
1.方法介绍
网络处理+后处理
2.网络介绍
1）由2个模块组成，line pre...]]></description><category></category></item><item><title><![CDATA[3D-LaneNet: End-to-End 3D Multiple Lane Detection]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/105443798</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/105443798</guid><author>xumi13</author><pubDate>Fri, 10 Apr 2020 22:35:19 +0800</pubDate><description><![CDATA[一.概述
时间：2018.11
作者：Noa Garnett
机构：General Motors Israel（通用汽车）
内容：基于前视图能够预测road coord 的3D lane，输出的是车道线3维曲线，即道路平面也考虑了。且不约束车道宽度或者已建图的环境。
3D lanenet：intra-network IPM + anchor-based lane representation。
数...]]></description><category></category></item><item><title><![CDATA[ubuntu安装tips——字体更新]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/105309880</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/105309880</guid><author>xumi13</author><pubDate>Sat, 04 Apr 2020 14:20:26 +0800</pubDate><description><![CDATA[由于母语是中文，建议安装完ubuntu系统后，把字体也更新一下：
1.准备字体：
1).从windows系统中拷贝过来：（推荐）

C:\Windons\Fonts

2).下载缺失的字体文件，然后复制到Linux系统中的/usr/share/fonts文件夹中。

国外下载地址：
https://www.dropbox.com/s/lfy4hvq95ilwyw5/wps_symbol_fonts...]]></description><category></category></item><item><title><![CDATA[Stacked Hourglass Networks for Human Pose Estimation]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/105185221</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/105185221</guid><author>xumi13</author><pubDate>Sun, 29 Mar 2020 20:20:31 +0800</pubDate><description><![CDATA[一.概述
1.paper: Stacked Hourglass Networks for Human Pose Estimation
2.code: http://www-personal.umich.edu/~alnewell/pose
3.Time: 201603
4.task: human pose estimation
5.eval_dataset: FLIC and MPII bench...]]></description><category></category></item><item><title><![CDATA[Key Points Estimation and Point Instance Segmentation Approach for Lane Detection]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/105181405</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/105181405</guid><author>xumi13</author><pubDate>Sun, 29 Mar 2020 16:47:46 +0800</pubDate><description><![CDATA[一.概述
paper: Key Points Estimation and Point Instance Segmentation Approach for Lane Detection
code: https://github.com/koyeongmin/PINet
dataset: Tusimple-dataset
Advantage:


The arbitrary number of l...]]></description><category></category></item><item><title><![CDATA[python2.7 install scikit-image]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/104605881</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/104605881</guid><author>xumi13</author><pubDate>Mon, 02 Mar 2020 10:00:08 +0800</pubDate><description><![CDATA[python2.7
1）pip install scikit-image
如果不行：
2）pip install scikit-image==0.14.1
如果还不行：
3)
pip install --user networkx==2.2
pip install --user dask==0.19.1
pip install --user PyWavelets==1.0
pip install ...]]></description><category></category></item><item><title><![CDATA[gitlab提交修改和更新到gitlab仓库步骤]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/104605822</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/104605822</guid><author>xumi13</author><pubDate>Mon, 02 Mar 2020 09:58:02 +0800</pubDate><description><![CDATA[gitlab提交修改和更新到gitlab仓库步骤：
1）git pull origin master
2）git status
如果有冲突，然后，用vs code打开这个本地仓库文件目录
将不是自己任务的代码选择远程同步到本地，自己的选择保存本地文件
再：git status
如果还有冲突的文件，不是需要管的，可以删除这个冲突：git checkout –  冲突文件名称
git status
3...]]></description><category></category></item><item><title><![CDATA[arc diff 和 git上传其他名称的分支]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/104605751</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/104605751</guid><author>xumi13</author><pubDate>Mon, 02 Mar 2020 09:54:01 +0800</pubDate><description><![CDATA[ignore一些文件
上传前，将一些大文件ignore：
https://github.com/github/gitignore选择指定格式的ignore文件，例如python,选择python.gitignore，复制内容到本地，本地文件夹内创建一个.gitignore文件。
上传其他名称的分支
在一个仓库中存入与仓库名称不一样的文件夹，作为分支。例如上传check_J2_31到其他名称的rep...]]></description><category></category></item><item><title><![CDATA[Deeplab v3+论文笔记]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/104602133</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/104602133</guid><author>xumi13</author><pubDate>Sun, 01 Mar 2020 22:40:27 +0800</pubDate><description><![CDATA[Deeplab v3+
Deeplab v3+的结构

1) Spatial pyramid pooling

encode multi-scale contextual  information by probing the incoming features.
2) encode-decoder structure
capture shaper object boundaries by gra...]]></description><category></category></item><item><title><![CDATA[空洞卷积Atrous convolution和ASPP]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/104584009</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/104584009</guid><author>xumi13</author><pubDate>Sat, 29 Feb 2020 22:44:00 +0800</pubDate><description><![CDATA[1.Atrous Convolution

from: https://towardsdatascience.com/review-deeplabv1-deeplabv2-atrous-convolution-semantic-segmentation-b51c5fbde92d

The term “Atrous” indeed comes from French “à trous” meanin...]]></description><category></category></item><item><title><![CDATA[python编程规范-google]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/104575514</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/104575514</guid><author>xumi13</author><pubDate>Sat, 29 Feb 2020 15:43:15 +0800</pubDate><description><![CDATA[https://zh-google-styleguide.readthedocs.io/en/latest/google-python-styleguide/python_language_rules/#lint
一.语法规范
1）对你的代码使用pylint；
2）仅对包和模块使用导入
使用 import x 来导入包和模块.
使用 from x import y , 其中x是包前缀, y是不带前...]]></description><category></category></item><item><title><![CDATA[ubuntu解压文件名称乱码方法]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/104114499</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/104114499</guid><author>xumi13</author><pubDate>Thu, 30 Jan 2020 13:37:10 +0800</pubDate><description><![CDATA[使用python解压文件
&lt;亲测可用，哈哈～～～&gt;
#!/usr/bin/env python
# -*- coding: utf-8 -*-

import os
import sys
import zipfile

#print "Processing File " + sys.argv[1]

file=zipfile.ZipFile(sys.argv[1],"r");
for ...]]></description><category></category></item><item><title><![CDATA[mxnet安装]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/101050056</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/101050056</guid><author>xumi13</author><pubDate>Fri, 20 Sep 2019 09:14:26 +0800</pubDate><description><![CDATA[1.安装过程
（建议在anaconda3环境下安装）
GPU版本：
根据自己的cuda版本选择相应的mxnet。
例如：cuda9.0环境：
pip install mxnet-cu90mkl==1.5.0

其中，‘90’表示cuda版本， ‘1.5.0’表示版本，‘mkl’表示mkl
CPU版本：
pip install mxnet-mkl==1.3.1

如果需要用到mxnet.viz可视化...]]></description><category></category></item><item><title><![CDATA[python实现：pdf 转换为 jpg/png/bmp]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/100539166</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/100539166</guid><author>xumi13</author><pubDate>Wed, 04 Sep 2019 14:56:46 +0800</pubDate><description><![CDATA[代码：
https://github.com/YaoLing13/pdf_to_image

]]></description><category></category></item><item><title><![CDATA[论文阅读- Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/99812921</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/99812921</guid><author>xumi13</author><pubDate>Sun, 01 Sep 2019 23:34:15 +0800</pubDate><description><![CDATA[Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer
论文地址：https://arxiv.org/abs/1612.03928
github地址：https://github.com/szagoruyko/atten...]]></description><category></category></item><item><title><![CDATA[分屏拔掉显示屏看不见窗口解决办法]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/100189605</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/100189605</guid><author>xumi13</author><pubDate>Sun, 01 Sep 2019 23:22:12 +0800</pubDate><description><![CDATA[启动该软件
按Alt+空格，然后按M(也可以鼠标选择” 移到 ”)
移到鼠标，即可看见软件界面，拖到本屏幕即可。

]]></description><category></category></item><item><title><![CDATA[车道线-论文阅读: Learning Lightweight Lane Detection CNNs by Self Attention Distillation]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/100185065</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/100185065</guid><author>xumi13</author><pubDate>Sun, 01 Sep 2019 17:54:07 +0800</pubDate><description><![CDATA[ICCV2019
code: https://github.com/cardwing/Codes-for-Lane-Detection
paper:  https://arxiv.org/abs/1908.00821
Abstract

present a novel knowledge distillation approach, i.e., Self Attention Distillatio...]]></description><category></category></item><item><title><![CDATA[论文阅读：Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/99711104</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/99711104</guid><author>xumi13</author><pubDate>Sun, 18 Aug 2019 23:15:00 +0800</pubDate><description><![CDATA[论文阅读：Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
论文地址： https://arxiv.org/pdf/1706.02677
背景
1）larger networks and larger datasets need longer time for training; 
解决方法:分布式同步SGD，将SGD minib...]]></description><category></category></item><item><title><![CDATA[vim快捷键]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/73166719</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/73166719</guid><author>xumi13</author><pubDate>Tue, 13 Jun 2017 11:36:08 +0800</pubDate><description><![CDATA[vi（vim）是上Linux非常常用的编辑器，很多linux发行版都默认安装了vi(vim)。vi(vim)命令繁多但是如果使用灵活之后将会大大提高效率。vi是“visual interface”的缩写，vim是vi IMproved(增强版的vi)。在一般的系统管理维护中vi就够用，如果想使用代码加亮的话可以使用vim。vi有3个模式：插入模式、命令模式、低行模式1）插入模式：在此模式下可以输入字]]></description><category></category></item><item><title><![CDATA[解决方法：fatal error C1083: 无法打开预编译头文件:“Debug\Open.pch”: No such file or directory]]></title><link>https://blog.youkuaiyun.com/xumi13/article/details/68063167</link><guid>https://blog.youkuaiyun.com/xumi13/article/details/68063167</guid><author>xumi13</author><pubDate>Wed, 29 Mar 2017 12:27:30 +0800</pubDate><description><![CDATA[项目属性→链接器→常规→启用增量链接，选择“否” 
然后，项目属性→C/C++→调试信息格式，选择“程序数据库” 
最后，“生成”，清除解决方案，再：重新生成解决方案。 
完成！]]></description><category></category></item></channel></rss>