t***r
发帖数: 225 | 1 浙大的吴粉丝可以上场了！
发信人: tiqer (虎子-粥星星-小丁丁), 信区: Returnee
标 题: Re: 来看看浙大大牛吴朝晖的水平吧
发信站: BBS 未名空间站 (Wed Oct 28 08:51:39 2009, 美东)
索性就一揭到底吧，原以为吴大牛的文章不多，是做产业为主的。其实不然，用DBLP看
了一下，竟然217篇，确实到了大牛的数量极了。
仔细看了一下，原来如此，跟学化学的灌SCI差不多呀。
大牛还是有少于7~8篇，还发表在不错的journal（IEEE KDE）和conference (ICCV,
ECCV, CVPR)上的。剩余的一小部分是2流大多是3流和不入流，比如International
Conference on Computational Science(16)，SMC（14）。
在比较一个美国的中牛吧（美国人），DBLP99篇，10篇PAMI+IJCV，20篇tier-1
conference in Computer vision, machine learning; 20篇MICCAI，30篇ICRA+IROS；
基本80%是主流或tie |
|
l****g
发帖数: 761 | 2 National Univ of Singapore
Computer Science Department Conference Rankings
Some conferences accept multiple categories of papers. The rankings
below are for the most prestigious category of paper at a given
conference (e.g. refereed full paper track).
All other categories should be treated as "unranked".
AREA: Artificial Intelligence and Related Subjects
Rank 1:
AAAI: American Association for AI National Conference
CVPR: IEEE Conf on Comp Vision and Pattern Recognition
IJ |
|
w****a
发帖数: 186 | 3 很粗略的分类，欢迎补充，全当抛砖引玉。
最牛的Conference:
International Conference on Computer Vision (ICCV)
Computer Vision and Pattern Recognition (CVPR)
European Conference on Computer Vision (ECCV)
差一些：
International Conference on Image Processing (ICIP)
International Conference on Pattern Recognition (ICPR)
Asian Conference on Computer Vision (ACCV)
最牛的Journal:
International Journal of Computer Vision (IJCV)
IEEE Trans. on Pattern Analysis and Machine Intellgence (PAMI) |
|
c********r
发帖数: 3 | 4 我认为应该是:
For Conference:
First tier: ICCV CVPR
1.5 tier: ECCV, ACM Multimedia
2nd tier: ICPR, ICME
3rd tier: ACCV, ICIP
For Journal:
First tier: PAMI, IJCV
2nd tier: CVIU, CVGIP |
|
h*******o
发帖数: 95 | 5 感觉IJCV比PAMI牛吧？
IJCV和ICCV是一个系列， PAMI和CVPR是一个系列 |
|
n******y
发帖数: 7 | 6 everyone has different arguments in comparing cvpr/iccv, or ijcv/pami.
but i strongly argue that accv is better than icpr/icip...
ac** is not so bad, considering asia siggragh, acml, pakdd...
the world is changing |
|
c****m
发帖数: 179 | 7 大家基本公认 ICCV CVPR ECCV 是first tier了，甚至很多牛人的简历里都会特别注明
一下。
至于acm mm，我觉得和icpr icme，accv没啥区别。 |
|
m****s
发帖数: 402 | 8 BMVC是四大会议中资历最久的，CV原发研究近一半是发自英国，ICCV也是CV热起来后87年才开始由英美两边权威联手办的。
自从有了ICCV，BMVC在CV中老大的地位就逐渐退却，现在连最新的ECCV都有超过的趋势。但无论如何，做CV资历深点的都不能不说BMVC，CVPR是年轻PhD们发飚的场合，BMVC是老先生聚会的地儿。 |
|
g***i
发帖数: 50 | 9 Note:
This ranking is collected from Internet, only for reference, maybe not accurate and complete.
Some conferences accept multiple categories of papers. The rankings below are for the most prestigious category of paper at a given conference. All other categories should be treated as "unranked".
AREA: Artificial Intelligence and Related Subjects
Rank 1:
AAAI: American Association for AI National Conference
CVPR: IEEE Conf on Comp Vision and Pattern Recognition
IJCAI: In |
|
n*o
发帖数: 442 | 10 In high performance computing, operating system,
network, vision/image processing, and database fields:
(not a complete list, but listed conferences are
generally ranked right.)
update 1: the conferences suggested by Yepes and TurnTurnTurn added
update 2: ics should be 4 star instead of 3 star as clarified by fengeer
5 star: OSDI, SOSP, SIGCOMM, MobiCom, VLDB, SIGMOD,
ASPLOS, MobiHoc, CVPR,ICCV
4 star: InfoComm, ICDE, FAST, NSDI, SC, HotOS, ICDCS,
DSN, ICNP, HPDC, IPTPS, USENIX, |
|
T**********n
发帖数: 480 | 11 vision & image processing
5:CVPR,ICCV
4:ECCV
3:ICIP
呵呵 |
|
y***s
发帖数: 294 | 12 你做图形还是数据库？貌似图形？
这几个会怎么样？SIGKDD、CVPR、SIGPLAN。
INFOCOM不算很低，>15%，在网络圈子里算比较仁慈的了。
MobiHoc也是一个<10%的top conference.
一般来说，感觉是好的工作投了一圈top conference不中之后，投到第二年的
INFOCOM上去。由于收的文章特别多，而且比较黑，一般你的老板如果在TPC里
面有人说话，质量还过得去的文章中的机会比较大。
INFOCOM真的是越来越不上档次了，它的余威就是靠每年200多篇文章里面
还是有那么一些好文章在撑着。今年的INFOCOM，这里做网络的faculty居然全
部按兵不动，都没去参加，包括有papers发在上面的。其声誉形象可想而知。
为之一叹。 |
|
r*****y
发帖数: 507 | 13 another fact about siggraph (it says siggraph is not prestigious at before):
it is true that siggraph submission each year is increasing, but slowly and
steadily.
several years ago, it accepted about 60 paper per year, last two years, it
accepted about 80 papers, this year (2005), it is said it accepted about 100
papers out of a little more than 400 paper submission.
accepted ratio is going toward ICCV and CVPR now.
for
get
of
limitation.
graphics
siggraph.
part
them |
|
S*********r
发帖数: 4729 | 14 ？你怎么知道的？
今天review结果才出来，都还没给acceptance通知。你说的是cvpr吧。 |
|
S*********r
发帖数: 4729 | 15 我们组一般，不算有名。每年平均2－3篇cvpr,或者iccv还是有的。 |
|
p*********e
发帖数: 48 | 16 是啊，把我CVPR文章的余热贡献出来。因为是“余热”，所以两个给4的reviewers说不够
novel，只是我先前的工作的改进。你们组的review那么狠，不会是落在你们手里吧？ |
|
k*********t
发帖数: 2999 | 17 too many people are submitting papers.
for example, someone in this bbs gotmore than 1 paper in cvpr and still had
papers sent to iccv... Somehow I cannot believe that one can have so many
decent work at once. so... |
|
e******y
发帖数: 15 | 18 do you know a big bull water N papers in this year's CVPR, (N>4), and
still have M submissions to this year's ICCV.
Can not understand, most time is spent in writing papers.
Maybe bull is different to normal human. |
|
S*********r
发帖数: 4729 | 19 I totally agree with you, it is sometimes unusual or should be considered
bad to publish paper in a rush. I don't believe those people who can make so
many papers at a single year. For instance, last year, one HongKong prof had 7
paper in CVPR, and 3 are oral. Every paper has only two authors, by him and
his student. I don't know how they can make this happen. Just writing 7 papers
can take 2 months.
|
|
T**********n
发帖数: 480 | 20 CVPR是人家的山头吧
也无可厚非，这个圈子也不是很大，基本上看个题目就知道是哪里的作品了
不过7篇也得算是人品相当不错了，呵呵 |
|
k*********t
发帖数: 2999 | 21 hehe. I was not angry at all. I myself published
“余热” papers too. just not so efficient.
vision
是啊，把我CVPR文章的余热贡献出来。因为是“余热”，所以两个给4的reviewers说
吧
天 |
|
k*********t
发帖数: 2999 | 22 The problem is that too many people have entered such an area with only 硬骨头
...
vision
是啊，把我CVPR文章的余热贡献出来。因为是“余热”，所以两个给4的reviewers说不
吧？ |
|
k*********t
发帖数: 2999 | 23 A Chinese folk got 5 in CVPR 2005 (submitted more. 2 or so orals) and
submitted another 5 or so to ICCV. He is not even an assistant professor. He
must be really good to finish so many in a year. Very impressive.
啊
科
点
many |
|
r**t
发帖数: 937 | 24 道理很简单，学生多，好学生多，老板的paper就多。大牛学生一年贡献两篇CVPR，
小牛学生一篇，如果手下有那么四五个，嘿嘿....... |
|
l**l
发帖数: 860 | 25 大牛学生除了每年两片cvpr 是不是顺带还噼里啪啦的写一些其他文章?
(非一稿多投) |
|
k*********t
发帖数: 2999 | 26 "publishing" in cvpr/iccv ...
， |
|
m*******y
发帖数: 48 | 27 客气点说你是mean，不客气点说你是无知无畏，
有本事你也灌7篇junk到CVPR！包括三个oral。 |
|
m*******y
发帖数: 48 | 28 说实话这几年CVPR ICCV出了多少真正好的paper？大家水平都差不多。
大家都是在灌水而已，看人家灌得多了就出来jjww，有这功夫自己也灌去呀。 |
|
s*****s
发帖数: 27 | 29 this group will have more than 5 papers in cvpr+iccv this year again.
some of them are oral.
you can not simply say they are rubbish, what I know is that the students
in that group are really smart guys, they are not at usa only because
they do not have taken English exam like toefl and gre or do not have
very good gpa (but good gpa is not closely related with his research ability as
I see). |
|
r**t
发帖数: 937 | 30 area chairs always know who the reviewers are.
they know how to choose.
but, for conferences such as ICCV or CVPR, one slightly negative review
is enough to kill a paper, if the negative comments really make sense to
the chair. however, sometimes a ridiculous negative comment still gives
the chair an excuse to kill your paper :(
life is really tough !! |
|
O******Y
发帖数: 88 | 31 That was not normal. However, I have never heard any accepted papers in cvpr
or iccv got a "5".
without
114 |
|
O******Y
发帖数: 88 | 32 That is very true. Too many submissions for vision conferences. So it is too
difficult to increase the number of reviewers...
cvpr |
|
S*********r
发帖数: 4729 | 33 My paper this time got a 5, really don't understand what that guy is
expecting to see from vision research. Our nice idea was then killed without
mercy.
cvpr |
|
S*********r
发帖数: 4729 | 34 Just wrote the rebuttal, hope they don't feel our tone is too strong. We
only hope it is in next CVPR anyway.
without
in |
|
O******Y
发帖数: 88 | 35 According to my experience, it is not helpful to have a strong tone. For last
CVPR, the comments from the area chair clearly said he did not like the strong
tone in our rebuttal and so the rebuttal was not taken into account. So be
careful.
a |
|
S*********r
发帖数: 4729 | 36 To my knowledge, ICCV won't be so low this year too.
They will keep the acceptance rate at the same level, but increasing the
number of accepted paper. Just like what they did for CVPR 2005. |
|
r*****y
发帖数: 507 | 37 that's enough. :) don't always think about siggraph.
let we think about another way.
in graphics,
siggraph accepted 90 papers, and the top ten graphics conferences/symposium (including siggraph) totally accepted less than TOTAL 450 papers annually.
and look at other fields, for example, computer vision.
top (ICCV or EECV + CVPR) accepted more than 500 papers each year (about 100 oral totally), so their oral = siggraph papers.
and second tire vision conf (ICPR, ACCV, etc.) accpeted other more tha |
|
g********g
发帖数: 11 | 38 This depends on the area.
In some area such as vision, the poster paper has the same length as the
regular paper. Actually there is no difference when you submit a paper (like
to CVPR, ICCV etc.). |
|
|
w****a
发帖数: 186 | 40 iccv and cvpr are equally important, in my opinion... |
|
F**p
发帖数: 1046 | 41 maybe CVPR is hotter now. ICCV was used to very good. |
|
l******e
发帖数: 153 | 42 It's harder to say. CVPR gets higher and higher accept ratio ... |
|
b*******s
发帖数: 14 | 43 don't think so. if only consider computer vision,
ICCV and ECCV should be better than CVPR. |
|
c******m
发帖数: 98 | 44 cvpr好像正在扩招，质量大概会下降
iccv是single track吧，感觉好一点
Can not see you iccv presentation, 555 |
|
l******e
发帖数: 153 | 45 My rank is
CVPR,ICCV > ECCV > BMVC,WACV > ACCV,ICASSP,ICIP,ICPR |
|
l******e
发帖数: 153 | 46 I know ICASSP is good for speech. However, for vision, as far as I know, there
is a lot of vision guys publish papers there, and it seems quite easy (accept
rate is around 50, if not above).
I think you are right, I should also include SIGGRAPH, and NIPS, it should be
like
CVPR,ICCV,SIGGRAPH,NIPS > ECCV > BMVC,WACV > ACCV,ICASSP,ICIP,ICPR
BTW, I don't think image=vision, they have different focus, though overlapped
quite some. |
|
c*********g
发帖数: 154 | 47 可以看05年CVPR（或者是ICCV，记不清了）上的一篇基于快速level set算法来做跟踪
的论文，写得挺清楚 |
|
t*****g
发帖数: 200 | 48 老大推荐一个最快给结果的journal 好么
pattern recognition
CVIU
IJCV
IEEE transaction SMA
IEEE transaction PAMI
IEEE transaction IP
IEEE transaction .....
btw 我觉得 cvpr oral 比很多jrnl好啊,至少比pattern recognition好吧? |
|
t*****g
发帖数: 200 | 49 搞gesture recognition 的, cvpr oral 是为了解决 "junk gesture" 问题的,是一个
很大的问题,
review 非常好.
应该还比较sound, 也可以generalize, btw,所有的paper都自己高出来的,不过说了别人也不信,因为是烂笑,所以觉得都是老板给的idea.
我实在搞不懂journal有什么意义.... 特别对于我这种烂笑的...
看的 |
|
e*******e
发帖数: 1144 | 50 Just sample some research areas related to AI and machine learning:
1) Computer vision, pattern recognition in image or video (in CVPR, ICCV)
2) Data mining, e.g., data streams, graph mining, anomaly detection (in KDD)
3) Web mining, text mining, information retrieval (in SIGIR, WWW)
4) Natural language processing (in EMNLP, ACL)
In your opinion, which are more promising in the near future (<10 years)
1) For an industry career (more demands from companies)
2) For an academic career (more faculty |
|
