-
編號一
編號二
Programming with libpcap
Set up local LAMP
Turn CubieBoard to Access Point
Step 1:
Load wifi module:
modprobe bcmdhd op_mode=2
Step 2:
Install hostapd:
sudo apt-get install hostapd
configure hostapd:
/etc/hostapd/
linaro@cubietruck:~$ cat /etc/hostapd/hostapd.conf
interface=wlan0
ctrl_interface=/var/run/hostapd
ssid=ssidname
channel=10
Step 3:
Run hostapd
hostapd /etc/hostapd/hostapd.conf
Ref:
http://forum.cubietech.com/forum.php?mod=viewthread&tid=1737&highlight=cubietruck%2Bap
CubieTruck Config Access Point
References:
For CubieTruck:
http://forum.cubietech.com/forum.php?mod=viewthread&tid=1580&highlight=cubietruck%2Bap
http://forum.cubietech.com/forum.php?mod=viewthread&tid=1737&highlight=cubietruck%2Bap
http://www.cubieforums.com/index.php/topic,1133.msg6743.html#msg6743
Scenario: Cubietruck with benn’s system using 3.4.61+ kernel
Problem: Unable to configure hostapd using bcmdhd for wifi. Tried standard hostapd.conf but my Tablet cannot see the device
iwlist wlan0 scan on ct shows the external wifi router.
linaro@cubietruck:~$ sudo iwlist wlan0 scan
wlan0 Scan completed :
Cell 01 – Address: 00:1E:40:06:C5:F8
ESSID:”jpmhome” <– That is my router
Mode:Managed
Frequency:2.462 GHz (Channel 11)
Quality:5/5 Signal level:-23 dBm Noise level:-92 dBm
IE: IEEE 802.11i/WPA2 Version 1
Group Cipher : WEP-40
Pairwise Ciphers (1) : CCMP
Authentication Suites (1) : PSK
Encryption key:on
Bit Rates:1 Mb/s; 2 Mb/s; 5.5 Mb/s; 11 Mb/s; 18 Mb/s
24 Mb/s; 36 Mb/s; 54 Mb/s; 6 Mb/s; 9 Mb/s
12 Mb/s; 48 Mb/s
linaro@cubietruck:~$ sudo iwconfig
lo no wireless extensions.
tunl0 no wireless extensions.
wlan0 IEEE 802.11 ESSID:””
Mode:Master Frequency:2.412 GHz Access Point: Not-Associated
Bit Rate:65 Mb/s Tx-Power:32 dBm
Retry min limit:7 RTS thr:off Fragment thr:off
Power Managementmode:All packets received
Link Quality=5/5 Signal level=0 dBm Noise level=0 dBm
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc:0 Missed beacon:0
eth0 no wireless extensions.
p2p0 IEEE 802.11bgn ESSID:off/any
Mode:Managed Access Point: Not-Associated Tx-Power=1496 dBm
Retry long limit:7 RTS thr:off Fragment thr:off
Encryption key:off
Power Management:on
linaro@cubietruck:~$ cat /etc/hostapd/hostapd.conf
interface=wlan0
ctrl_interface=/var/run/hostapd
ssid=skylab
channel=1
P.S: The Cubieboard2 worked well as an AccessPoint using Dongle supplied by Cubie.
Appreciate a way to solve this.
Joseph Mathew
Modified:
Solved it by modifying bcmdhd module load :
bcmdhd op_mode=2
Now the EESID is set and can access from another Tablet.
linaro@cubietruck:~$ sudo iwconfig
wlan0 IEEE 802.11 ESSID:”skylab”
Mode:Master Frequency:2.412 GHz Access Point: 02:1A:11:FE:34:96
Bit Rate=65 Mb/s Tx-Power:32 dBm
Retry min limit:7 RTS thr:off Fragment thr:off
Power Managementmode:All packets received
Thanks !!
Cubietruck is a great board!! and I am all set to use that in my School Education project: SkyLab
Whether your wireless card support Access Point mode.
First thing to be done is perform the test whether your wireless card support going into wireless access point mode. As told earlier following test is for mac80211 framework based driver.
Install iw
& execute following
sudo aptitude install iw
iw list
Look for supported interface section, where it should be a entry called AP
like below
Supported interface modes:
* IBSS
* managed
* AP
* AP/VLAN
* monitor
* mesh point
If your driver doesn’t shows this AP, It doesn’t mean it can’t create wireless hotspot. But those cards aren’t in scope of this tutorial. For more tests follow ubuntu documentation on master mode.
The setup is divided in three sections,
- Setup & host a wireless network
- IP address setup
- Internet sharing
1.Setup and host a network
- Software required: hostapd (install it)
- Press alt + F2 and type
gksu gedit
& press enter. We are going to edit a lot of files. - In gedit, press ctrl+o, ctrl+l & paste it in location box
/etc/hostapd/hostapd.conf
. Press Enter. - Paste the following code,
interface=wlan0 driver=nl80211 ssid=test hw_mode=g channel=1 macaddr_acl=0 auth_algs=1 ignore_broadcast_ssid=0 wpa=3 wpa_passphrase=1234567890 wpa_key_mgmt=WPA-PSK wpa_pairwise=TKIP rsn_pairwise=CCMP
Changes you need to do:
- Change
interface=wlan0
to your wireless card name. (If you have one wireless card it should be wlan0) ssid=test
.test
is the name of your hosted network.wpa_passphrase=1234567890
,1234567890
is the password of your network.
The configuration above creates a wpa & wpa2 enabled access point in g
mode. A more detailed instruction to build configuration file can be found here
Now start the hostapd
. Edit the file /etc/default/hostapd
and modify the line of DAEMON_CONF like this:
DAEMON_CONF="/etc/hostapd/hostapd.conf"
Then start the hostapd
service using the following command,
sudo service hostapd start
It should start a wireless network. In your mobile device now you can see a wireless network and can authenticate. But the device won’t get IP address. Stop it with the command sudo service hostapd stop
If you get any error, possibly your card doesn’t support
g
mode. Try with other >modes. Guide
Part 2: Set up DHCP server for IP address management
Install isc-dhcp-server
Edit the file /etc/default/isc-dhcp-server
and set INTERFACES like this:
INTERFACES="wlan0"
In gedit, press Ctrl+O, in location box paste /etc/dhcp/dhcpd.conf
Find (ctrl+F) below lines and put # before it. It should look like after editing
# option definitions common to all supported networks…
#option domain-name “example.org”;
#option domain-name-servers ns1.example.org, ns2.example.org;
Again comment out following lines too
#default-lease-time 600;
#max-lease-time 7200;
Add following lines at end
subnet 10.10.0.0 netmask 255.255.255.0 {
range 10.10.0.2 10.10.0.16;
option domain-name-servers 8.8.4.4, 208.67.222.222;
option routers 10.10.0.1;
}
Range describe how long the address pool will be. you need to adjust subnet value also. This config can give IP up to 15 devices
Again press Cctrl+O in gedit and paste following in location bar /etc/network/interfaces
, Add below
auto wlan0
iface wlan0 inet static
address 10.10.0.1
netmask 255.255.255.0
wlan0
is your wireless interface. Change it accordingly.
Note: After reboot the wireless will be shown as not managed. So you can’t use any other wi-fi network. To get wireless with normal behaviour, put # before those newly added line and execute
sudo start networking
now run
sudo service isc-dhcp-server start
sudo service hostapd start
At this point , your mobile device will see a network, authenticate it & after authentication it will get ip address something like 10.10.0.2
.
Setup internet connection settings
For Internet connection sharing we need ip forwarding
and ip masquerading
. Enable ip forwarding : execute
echo 1| sudo tee /proc/sys/net/ipv4/ip_forward
Now say you are using to a dial up/usb modem connection to connect to INTERNET. You need to get the logical interface name. For that execute ifconfig
or ip address
For dialup/usb modem: it should be ppp0
. If you want to share Ethernet connection you should useethX
where X is your ethernet device number. If you are connecting to internet through an android device with USB tethering, then the interface name should be usb0
.
Now once you get the interface name execute sudo iptables -t nat -A POSTROUTING -s 10.10.0.0/16 -o ppp0 -j MASQUERADE
The ppp0
in above command is the interface whose internet connection you are sharing over wireless.
If you have edited upto /etc/network/interfaces
you can use this script to start the service. Edit it if you are not sharing ppp0
. if you are using script and want to stop the server, use sudo killall hostapd
A blog that is mostly a summary of what has been described here:http://dashohoxha.blogspot.com/2013/06/how-to-setup-wifi-access-point-on-ubuntu.html
A script that tries to automate the steps described here: https://gist.github.com/dashohoxha/5767262
Please feel free to report or fix any bugs that you notice.
Great helps from :
转载 — Ubuntu on ThinkPad T430s (and W530)
About half of my team including me received refreshed laptops a few weeks ago. We made the push for Apple MacBooks again. Like the last time we have new Lenovo ThinkPads.
The team ordered Core i7 T430s ThinkPads while one team member, Jim, ordered the Core i7 W530. In retrospect I should have gone with the X230, but I’m stuck with what I got. I pulled the factory hard drive for a super fast SSD prior to powering up for the first time. I also maxed out the RAM with 16GB. My initial Windows 7 install was lackluster. Jim went straight to Ubuntu 12.04. He struggled with it a bit. Oddly it was roughly as much as I was struggling with Windows 7.
I decided to follow him down the Ubuntu 12.04 64-bit path. Here I’ll try to document what we’re doing on these two platforms. If this is useful to you and/or you have changes to submit, please comment here.
First off, a lot of things worked out of the box: wired network, wireless network, camera, sound, microphone, bluetooth, keyboard backlight, keyboard light, sleep, power management, screen brightness, driving two displays, touchpad, usb3.
UPDATED: Hibernation works. Hybrid sleep, too!
Things known not to work yet: fingerprint reader, Ricoh MMC/SD reader, 3G/4G activation (doesn’t work on Windows, either)
Not yet tested: smartcard reader, thunderbolt, 3G/4G, hibernation (I read it’s broken)
Jim and I have the advantage of our old laptops. They still work so we can experiment a bit without it impacting our ability to work.
BIOS
We did a few BIOS changes.
Under Config, Network make sure that Wake on LAN is disabled.
Under Config, Display make sure the Graphics Device is “Integrated Graphics” unless you plan on a three headed display.
Under Config, Power enable “Power On with AC Attach”
Under Security, Virtualization make sure both options are enabled.
Ubuntu Install
The installer errored when I asked it to encrypt my home folder during install.
If you have an Internet connection go ahead and have it install updates during install. You’ll still want to check for updates after you reboot to complete the install.
Work Apps: Juniper NSM
We use Network Security-Manager and NetScreen Security Manager (NSM) to manage Juniper devices. They require the ia32-libs metapackage to run. Well, they don’t need all of them but I’m not identifying each and every library manually.
After installing ia32-libs, do the following:
sudo ln -s /lib/x86_64-linux-gnu/libc-2.15.so /lib/libc.so.6 sudo ln -s /usr/lib/x86_64-linux-gnu/libXp.so.6.2.0 /usr/lib/libXp.so.6
Work Apps: IBM Lotus Notes
While installing Notes in Ubuntu is an option, I decided to install it in a Windows VM guest. You’ll need the ia32-libs if you chose to run Notes in Ubuntu.
Work Apps: Juniper SA & Network Connect
We also use the Juniper SA for remote access, so we need Network Connect. I followed the instructionshere to get Juniper SA Network Connect working. I used the OpenJDK 7 JRE and icedtea 7 plugin for the browser.
UPDATE: I also installed …
sudo apt-get install zlib1g:1386
Work Apps: VMWare Workstation
We’re both evaluating VMWare Workstation for the few apps we need a real Windows instance for. Version 9 installed perfectly. Windows 7 and 8 both install just fine. Unity, the VMWare mechanism for making a guest app look like a host app, works well.
However, in Windows 8 the Windows/super key is indispensable. Ubuntu Unity wants to pop-up and overlay with keyboard shortcuts, obscuring the stuff behind it. If you have the CompizConfig Settings Manager (CCSM, not recommended by Ubuntu), select Ubuntu Unity Plugin, go to the “Experimental” tab and disable “Enable Shortcut Hints Overlay”.
UPDATE: I had to add myself to the floppy group to get at my CDROM drive. For some reason Ubuntu mounts it as /media/floppy0.
Power
I travel for work. If I carry this unfortunately heavier-than-my-last laptop I want to maximize the battery life when detached. I want strong performance when I’m powered.
In the Power settings under “On Battery Power” I suspend when inactive for 10 minutes, power off when power is critically low, and suspend when the lid is closed. When plugged in, I don’t suspend when inactive and do nothing when the lid is closed.
I install powertop, thinkfan, thinkpad-acpi-dkms, and ethtool. Note that powertop gives you the best information when you’re on battery.
From the Ubuntu Wiki:
Enable ALPM:
echo SATA_ALPM_ENABLE=true | sudo tee -a /etc/pm/config.d/sata_alpm
I installed thinkfan for improved cooling with some help from here:
echo options thinkpad_acpi fan_control=1 | sudo tee -a /etc/modprobe.d/thinkfan.conf
UPDATE: I’m now using thinkpad-acpi-dkms as thinkfan doesn’t seem to work any more. More information here.
I disabled Wake-on-LAN on my ethernet interfaces: go here and scroll down to “For Ubuntu 12.04 (and up) users …” and step 5.
UPDATE: The post is a little confusing and non-linear, so do the following:
sudo cp /usr/lib/pm-utils/power.d/disable_wol /etc/pm/power.d
vi /etc/pm/power.d/disable_wol
… and change line 14 to look like …
enable) ethtool -s "${d##*/}" wol d>/dev/null 2>&1;;
I remounted my root partition to turn off atime tracking:
sudo vi /etc/fstab
and add ‘noatime’ to make your root entry look like …
UUID=4ba02cd9-d3c4-4dd1-9a30-6535fcba5290 / ext4 noatime,errors=remount-ro 0 1
Display
Jim’s W530
Jim dove into driving two external displays and the built-in display with an extended desktop. He did a number of steps on his W530 from here.
sudo add-apt-repository ppa:ubuntu-x-swat/x-updates
sudo apt-get update
sudo apt-get install nvidia-current
sudo gedit /etc/default/grub
… and on the line with “quiet splash” add “nox2apic” after “quiet”. Then do a …
sudo update-grub
… and then reboot.
UPDATE: Enter the BIOS. Enable “Discrete Graphics” or “NVidia Optimus”. Jim has to log in twice to his desktop, which is odd.
Paul’s T430s
I am trying to get the T430s to drive the displays without making the GRUB changes. I know from the Windows 7 experience on this laptop that you have to use DisplayPort for both of the external displays to drive all three. When I did this with my initial install, I could run three displays. However, the would all blank out then restore every 30 seconds or so.
Digging around a bit I found an article that discussed the Intel 4000 graphics chipset. The recommendation is to upgrade the kernel to the version in 12.10. I decided to go straight to 12.10. Unfortunately my work on getting some of my security tools installed made the upgrade not work as it should have. I decided to reinstall fresh and immediately upgrade to 12.10.
I noticed an immediate improvement after rebooting. The login screens for the two displays I had connected – an external and the built-in display – ran the correct (or close to) resolution for both. I made the usual post-install display adjustments to turn off mirroring and drive the external display at it’s native resolution.
Bolstered by the good look I decided to plug my second external display into the other DisplayPort. The mouse remained on the first two displays but they were blank otherwise. I switched to a terminal (Ctrl-Alt-F1) to do a ‘sudo unity –reset’ which unfortunately did nothing. The option is depreciated. However, the terminal did display on all three monitors. I rebooted, and all three displays showed the login screen. I logged in, and all three had the desktop in mirror mode.
I turned off “Mirror Displays” and they all show. When I tweak the external displays’ resolution I find I can’t quite drive them at their recommended resolution – 1680×1050 for one and 1920×1080 for the other. Even redoing the displays at a lower resolution while maintaining the aspect ratio didn’t help. I replaced the larger display with one that matches the other one.
After more experimentation I found that one of the DisplayPort to DVI cables is bad. I’ll have to test this out in the office on Monday.
UPDATE 25 Oct 12: I followed Jim’s steps above on 12.04 and it worked! For a while it worked, that is, and without the change to grub. I undocked the laptop and re-docked but the displays wouldn’t restore properly. Then they would black out. No amount of rebooting or powering off would bring it back.
UPDATE 31 OCT 12: I made changes to grub detailed here. These changes stabilized things.
Miscellaneous
Emacs
I’m an Emacs fan, especially of version 24. Follow the instructions here. Unfortunately Emacs 24 isn’t available by default.
Firefox
Of course I followed my own advice and set up my Widescreen Firefox. One note is that under Ubuntu with the default Ubuntu add-ons installed you do not need the “Hide Forward/Back Buttons When not Needed“ Stylish script.
I keep looking for a way to make the Alt Text in Firefox, where you mouse over a link in a web page and the full URL is displayed above the status bar at the bottom of the Firefox window, much lighter than it is by default. Out of the box it is black text on a dark gray background. If anyone has a solution, please share.
Other
I’m a big EverNote user. They do not have a Linux client yet, so I wanted to use NixNote. However, it required 32-bit Java. I’m trying EverPad instead.
To keep Ubuntu from popping up a window to upgrade,
gconftool-2 --set --type bool /apps/update-notifier/auto_launch false
I also use aptitude, synergy, and a bunch of other tools. I write about them from time to time.
Major Tasks in 2013
Job application:
1. Imperial College London Funding application
2. Oxford Funding Application
3. ETH Zurich
4. EPFL
5. KTH
6. MSRA
7. U.S Company
Conferences papers
1. ICCV – April 12th
2. BMVC – April 24th
3. CVPR 2014 – December
4. ICPR 2014 – December
[zz]计算机视觉的会议与专家
computer vision(cv) 存在ICCV/CVPR/ECCV三个顶级会议, 它们档次差不多,都应该在一流会议行列, 没有必要给个高下. 有些us的人认为ICCV/CVPR略好于ECCV,而欧洲人大都认为ICCV/ECCV略好于CVPR, 某些英国的人甚至认为BMVC好于CVPR.简言之, 三个会议差不多, 各有侧重和偏好.三者乃cv领域的旗舰和风向标,其oral paper (包括best paper) 代表当年度cv的最高水准, 在此引用Harry Shum的一句话, 想知道某个领域在做些什么, 找最近几年此领域的proceeding看看就知道了. ICCV/CVPR由IEEE Computer Society牵头组织, ECCV好像没有专门负责的组织. CVPR每年(除2002年)都在美国开, ECCV每两年开一次,仅限欧洲, ICCV也是每两年一次, 各洲轮值. 基本可以保证每年有两个会议开, 这样研究者就有两次跻身牛会的机会.
就录取率而言, 三会都有波动. 如ICCV2001录取率>30%, 且出现两个人(华人)各有三篇第一作者的paper的情况, 这在顶级牛会是不常见的 (灌水嫌疑). 但是, ICCV2003, 2005两次录取率都很低, 大约20%左右. ECCV也是类似规律, 在2004年以前都是>30%, 2006年降低到20%左右. CVPR的录取率近年来一直偏高, 从2004年开始一直都在[25%,30%].最近一次CVPR2006是28.1%, CVPR2007还不知道统计数据. 笔者猜测为了维持录取paper的绝对数量, 当submission少的时候录取率偏高, 反之偏低, 近几年三大会议的投稿数量全部超过1000, 相对2000年前, 三会录取率均大幅度降低, 最大幅度50%->20%. 对录取率走势感兴趣的朋友, 可参考http://vrlab.epfl.ch/~ulicny/statistics/ (CVPR2004的数据是错的),
http://www.adaptivebox.net/research/bookmark/CICON_stat.html .
显然, 投入cv的人越来越多,这个领域也是越来越大, 这点颇不似machine learning一直奉行愚蠢的小圈子主义. 另外一点值得注意, ICCV/ECCV只收vision相关的topic, 而cvpr会收少量的pattern recognition paper, 如fingerprint等, 但是不收和image/video完全不占边的pr paper,如speech recognition等. 我一个朋友曾经review过一篇投往CVPR的speech的paper, 三个reviewer一致拒绝, 其中一个reviewer搞笑的指出, 你这篇paper应该是投ICASSP被据而转投CVPR的. 就topic而言, CVPR涵盖最广. 还有一个没有验证过的原因导致CVPR录取率高: 很多us的researcher不愿意或没有足够的经费到us以外的地方开会, 故CVPR会优先接收很多来自us的paper (让大家都happy).
以上对三会的分析对我们投paper是很有指导作用的. 目前的research我想绝大部分还是纸上谈兵, 必经 read paper -> write paper -> publish paper -> publish paper on top conferences and journals流程. 故了解投paper的一些基本技巧, 掌握领域的走向和热点, 是非常必要的. 避免做无用功,选择切合的topic, 改善presentation, 注意格式 (遵守规定的模板), 我想这是很多新手需要注意的问题. 如ICCV2007明文规定不写summary page直接reject, 但是仍然有人忽视, 这是相当不值得的.paper毕竟是死的, 写paper的人才是活的. 那么我现在研究一下cv圈的格局, 按师承关系,
借鉴前人, 我总结a tree stucture of cv guys.
David Marr
—–>Shimon Ullman (Weizmann)
—–>Eric Grimson (MIT)
—–>Daniel Huttenlocher (Cornell)
—–>Pedro Felzenszwalb (Chicago)
Thomas Binford (Stanford)
—–>David Lowe (UBC)
—–>Jitendra Malik (UC Berkeley)
—–>Pietro Perona (Caltech)
—–>Stefano Soatto (UCLA)
—–>Fei-Fei Li (Princeton)
—–>Jianbo Shi (UPenn)
—–>Yizhou Yu (UIUC)
—–>Christoph Bregler (NYU)
—–>Serge Belongie (UCSD)
—–>Alyosha Efros (CMU)
Andrew Blake (Microsoft Research Cambridge)
—–>Andrew Zisserman (Oxford)
—–>Andrew Fitzgibbon (Microsoft Research Cambridge)
—–>Roberto Cipolla (Cambridge)
—–>Alan Yuille (UCLA)
(UK这个学派的师承关系不太清楚, 这是我听说加上自己猜测的. 事实上, 几个非常优秀的researcher如Vladimir Kolmogorov虽然不是Andrew Blake的学生, 但是也属于这个学派. )
Thomas Huang (UIUC)
—–>Yong Rui (Microsoft Research Redmond)
—–>Nebojsa Jojic (Microsoft Research Redmond)
—–>Ying Wu (Northwestern University)
—–>Hai Tao (UCSC)
—–>Yuncai Liu (SJTU)
(Huang这个系的人太多, 而且很怪的是, UIUC的web上信息不全, 在此仅列出我知道的.)此外, 还有Takeo Kanade等非常有名的大牛, 囿于篇幅, 不一一列举. 从上得知, 加州派基本占了cv的半壁江山. 最近几年, 特别活跃的cv guys是USA
Jitendra Malik, UC Berkeley
Pietro Perona, Caltech
Serge Belongie, UCSD
Jianbo Shi, UPenn
Stefano Soatto, UCLA
Fei-Fei Li, Princeton
William Freeman, MIT
Trevor Darrell, MIT
Simon Baker, CMU
Yanxi Liu, CMU
Songchun Zhu, UCLA
Alan Yuille, UCLA
Yi Ma, UIUC
Michael Black, Brown
Carlo Tomasi, Duke
Ramin Zabih, Cornell
Shree Nayar, Columbia
Rama Chellappa, Maryland
Steve Seitz, University of Washington
Europe
Andrew Zisserman, Oxford, UK
Andrew Fitzgibbon, Microsoft Research Cambridge, UK
Roberto Cipolla, Cambridge, UK
Jean Ponce, INRIA, France
Cordelia Schmid, INRIA, France
Bill Triggs, LEAR, France
Yair Weiss, Hebrew University, Israel
Anat Levin, Hebrew University, Israel
Michal Irani, Weizmann, Israel
Luc van Gool, University of Leuven/ETH Zurich, Czechic
China
Harry Shum, MSRA
Xiaoou Tang, MSRA/CUHK
Jian Sun, MSRA
Steve Lin, MSRA
Yasuyuki Matsushita, MSRA
Zhouchen Lin, MSRA
Long Quan, HKUST
Chi-Keung Tang, HKUST
就写到这, 希望这些信息对大家有用.
补充一下那个tree,一个法国大牛
Olivier Faugeras
—-Ponce UIUC
—lazebnik
—-Zhengyou Zhang MSR
—-Martial Hebert CMU
Mit AI lab
poggio
—Oliva
—serre
Freeman 80年代还来太原理工扶贫了
—Y. Weiss
—– Levin
—Antonio Torralba (research scientist)
Trevor Darrell
—Grauman
补充一下
Zisserman还有一个不错的学生
lifeifei的合作者Rob Fergus
按研究方向分分,应该更合理一些。
现在计算机视觉,计算机图形图像,机器学习开始融合到一起了吧。
J. Malik,Zhu Songchu偏segmentation;
D. Lowe, S. Ullman, Poggio 偏于从生物视觉的启发来研究视觉;
Zisserman, Schmid, Lowe研究局部特征;
Luc Van Goo, Long Quanl三维重建;
Perona, Li Feife, Freeman视觉i学习, 物体分类;
还有运动分析,视觉跟踪,纹理分析………….
MIT的Brain & Cognitive Science Dept和CSAIL里面聚集了一帮人,有的作low level有的作mid level to high level的。他们的工作是值得关注的。当然说视觉还是要从伟大的David Marr开始。Tomaso Poggio, Richard Whitman是Marr的同事,传承了其理念,一直往下走。Poggio最近几年比较重点的工作放在他那个hierarchical model上。T. Poggio的第一个PhD学生是Christof Koch (kLab at Caltech)。哦,顺便说一下Koch的另外一个导师是Valentino Breitenberg——同样是影响了一个时代的大人物。Koch研究重点兴趣在consciousness上,在Nature上的很多文章体现了他的研究思想。不过他们也做不少初级的视觉问题,诸如attention。
Koch比较知名的弟子比如Laurent Itti和Li Feifei。
Richard Whitman 年纪比较大了,个人不是很关注他现在做的东西。不过他所在的Perceptual Science Group,是一个非常有影响力的地方,这个组其他大家比较熟悉的老师有Aude Oliva和EH Adelson。Adelson最著名的一个事儿是色彩恒常相关的视错觉,93年发在Science上的那篇。关于Oliva,前面的帖子错了,她不是Poggio的学生,这家伙和Torralba是老乡,同在法国念书,主要从心理学那条路子开始做,成名之役是hybrid image,和Antonio Torrralba一起搞的。这个Hybrid Image 其实80年代就有了,但是最开始从心理学方向上探讨,没有非常有影响力的文章。后来开始靠谱作自然图像统计,得到Gist theory,当然这个illusion本身后来转投SIGGRAPH,其影响是深远的。嗯,这个和CV关系不大。Perceptual Science Group出了不少牛人,他们的alumni list可谓超豪华阵容:Yair Weiss, Josh Tenenbaum, Pawan Sinha, Bill Freeman……
这一派的工作跟我比较相关,大概八卦的关系也能总结出一些。
希望对大家有所帮助
Van 那个组也做局部特征和物体识别分类。现在Olivier Faugeras也从三维重建转到计算认知学了。ctozlm, 现在不研究局部特征的很少啊,因为局部特征方法克服了以前方法的很多缺陷。
局部特征方法本就是从三维重建发展出来的。所以他们研究局部特征也就不奇怪了。INRIA 的 Faugeras 是 IJCV 的 主编,在欧洲 Computer Vision 届是老大,他的弟子Ayach 现在 Medical Image Analysis领域有超过其师之势。
关于computer vision的会议及vision guys – 研究探讨 – 模式识别爱好者论坛模式识别,机器学习,人工智能,人脸识别,人脸检测 – Powered by Discuz!。发表于 2007-5-24 00:45
ter=type&typeid=2″>http://prfans.com/forum/viewthread.php?tid=65&extra=page=1&filter=type&typeid=2
计算机视觉牛人(转载)
http://blog.shamoxia.com/html/y2009/110.html
http://www.cnblogs.com/frankman/archive/2008/10/06/1304979.html
以下是我经常浏览的网络资源,关注大牛的网页,比上学校数据库资源更精、更有启发性。排名不分先后,呵呵~~~
(1)微软公司的文献:http://research.microsoft.com/research/pubs
(2)微软亚洲研究院:http://research.microsoft.com/asia/,值得关注Harry Shum, Jian Sun, Steven Lin, Long Quan(兼职HKUST)etc.
(3)瑞典隆德大学数学系视觉组:http://www.maths.lth.se/matematiklth/personal/andersp/
感觉国外搞视觉的好多是数学系出身,大约做计算机视觉对数学要求很高吧。
(4)澳大利亚国立大学:http://users.rsise.anu.edu.au/~hartley/
(5)美国北卡大学:http://www.cs.unc.edu/~marc/
(6)加州大学伯克利分校David A. Forsyth:http://www.cs.berkeley.edu/~daf/
(7)CMU的视觉组:http://www.cs.cmu.edu/~cil/vision.html
著名的有Tomasi, Kanade等,CMU不愧是美国计算机牛校,仅视觉就好猛。
(8)法国INRIA:http://www-sop.inria.fr/odyssee/team/
由Olivier.Faugeras领衔的牛人众多。
(9)英国牛津的A.Zisserman:http://www.robots.ox.ac.uk/~az/
(10)比利时鲁汶大学的L.Van Gool: www.esat.kuleuven.ac.be/psi/visics/
据说在这个只有中国一个小镇大小的地方的鲁汶大学在欧洲排行top10,名列世界top100,还出了几个诺贝尔奖,视觉研究也很牛,真是让Chinese汗颜啊!
http://www.sciencenet.cn/m/user_content.aspx?id=19093
转自:http://www.shamoxia.com/html/y2009/286.html
以下链接是本人整理的关于计算机视觉(ComputerVision, CV)相关领域的网站链接,其中有CV牛人的主页,CV研究小组的主页,CV领域的paper,代码,CV领域的最新动态,国内的应用情况等等。打算从事这个行业或者刚入门的朋友可以多关注这些网站,多了解一些CV的具体应用。搞研究的朋友也可以从中了解到很多牛人的研究动态、招生情况等。
总之,我认为,知识只有分享才能产生更大的价值,真诚希望下面的链接能对朋友们有所帮助。
(1)googleResearch; http://research.google.com/index.html
(2)MIT博士,汤晓欧学生林达华; http://people.csail.mit.edu/dhlin/index.html
(3)MIT博士后Douglas Lanman; http://web.media.mit.edu/~dlanman/
(4)opencv中文网站; http://www.opencv.org.cn/index.php/首页
(5)Stanford大学vision实验室; http://vision.stanford.edu/research.html
(6)Stanford大学博士崔靖宇; http://www.stanford.edu/~jycui/
(7)UCLA教授朱松纯; http://www.stat.ucla.edu/~sczhu/
(8)中国人工智能网; http://www.chinaai.org/
(9)中国视觉网; http://www.china-vision.net/
(10)中科院自动化所; http://www.ia.cas.cn/
(11)中科院自动化所李子青研究员; http://www.cbsr.ia.ac.cn/users/szli/
(12)中科院计算所山世光研究员; http://www.jdl.ac.cn/user/sgshan/
(13)人脸识别主页; http://www.face-rec.org/
(14)加州大学伯克利分校CV小组; http://www.eecs.berkeley.edu/Research/Projects/CS/vision/
(15)南加州大学CV实验室; http://iris.usc.edu/USC-Computer-Vision.html
(16)卡内基梅隆大学CV主页; http://www-2.cs.cmu.edu/afs/cs/project/ … ision.html
(17)微软CV研究员Richard Szeliski; http://research.microsoft.com/en-us/um/people/szeliski/
(18)微软亚洲研究院计算机视觉研究组; http://research.microsoft.com/en-us/groups/vc/
(19)微软剑桥研究院ML与CV研究组; http://research.microsoft.com/en-us/gro … fault.aspx
(20)研学论坛; http://bbs.matwav.com/
(21)美国Rutgers大学助理教授刘青山; http://www.research.rutgers.edu/~qsliu/
(22)计算机视觉最新资讯网; http://www.cvchina.info/
(23)运动检测、阴影、跟踪的测试视频下载; http://apps.hi.baidu.com/share/detail/18903287
(24)香港中文大学助理教授王晓刚; http://www.ee.cuhk.edu.hk/~xgwang/
(25)香港中文大学多媒体实验室(汤晓鸥); http://mmlab.ie.cuhk.edu.hk/
(26)U.C. San Diego. computer vision; http://vision.ucsd.edu/content/home
(27)CVonline; http://homepages.inf.ed.ac.uk/rbf/CVonline/
(28)computer vision software; http://peipa.essex.ac.uk/info/software.html
(29)Computer Vision Resource; http://www.cvpapers.com/
(30)computer vision research groups; http://peipa.essex.ac.uk/info/groups.html
(31)computer vision center; http://computervisioncentral.com/cvcnews
(32)浙江大学图像技术研究与应用(ITRA)团队:http://www.dvzju.com/
(33)自动识别网:http://www.autoid-china.com.cn/
(34)清华大学章毓晋教授:http://www.tsinghua.edu.cn/publish/ee/4157/2010/20101217173552339241557/20101217173552339241557_.html
(35)顶级民用机器人研究小组Porf.Gary领导的Willow Garage: http://www.willowgarage.com/
(36)上海交通大学图像处理与模式识别研究所:http://www.pami.sjtu.edu.cn/
(37)上海交通大学计算机视觉实验室刘允才教授:http://www.visionlab.sjtu.edu.cn/
(38)德克萨斯州大学奥斯汀分校助理教授Kristen Grauman :http://www.cs.utexas.edu/~grauman/
(39)清华大学电子工程系智能图文信息处理实验室(丁晓青教授):http://ocrserv.ee.tsinghua.edu.cn/auto/index.asp
(40)北京大学高文教授:http://www.jdl.ac.cn/htm-gaowen/
(41)清华大学艾海舟教授:http://media.cs.tsinghua.edu.cn/cn/aihz
(42)中科院生物识别与安全技术研究中心:http://www.cbsr.ia.ac.cn/china/index CH.asp
(43)瑞士巴塞尔大学 Thomas Vetter教授:http://informatik.unibas.ch/personen/vetter_t.html
(44)俄勒冈州立大学 Rob Hess博士:http://blogs.oregonstate.edu/hess/
(45)深圳大学于仕祺副教授:http://yushiqi.cn/
(46)西安交通大学人工智能与机器人研究所:http://www.aiar.xjtu.edu.cn/
(47)卡内基梅隆大学研究员Robert T. Collins: http://www.cs.cmu.edu/~rcollins/home.html#Background
(48)MIT博士Chris Stauffer: http://people.csail.mit.edu/stauffer/Home/index.php
(49)美国密歇根州立大学生物识别研究组(Anil K. Jain教授):http://www.cse.msu.edu/rgroups/biometrics/
(50)美国伊利诺伊州立大学Thomas S. Huang:http://www.beckman.illinois.edu/directory/t-huang1