Called to order at 9:11 EDT by Avi Kak.
#1. CVPR93 (Aggarwal, Aloimonos, Bolle). Between 440 and 450 people
and making too much money. Papers 450 papers accepted 89 and 90
Recognition for local arrangements (Terry Boult).
#2 Discussion (Kak): did we broaden the scope? After the last one
some felt that CVPR had been too narrow and some feel that it is
still too narrow. Open discussion on whether it is broad enough.
Time to discuss the issue for next meeting.
(AK:) should be sessions on Computer Vision and Image Processing and
Computer Vision and document analysis. (Haralick) perspective on
origin of CVPR from PRIP for the purpose to bring in Computer Vision
to balance Pattern Recognition and IP. Over the years, Computer
Vision has increased to the point where Computer Vision dominates
(where is PR?). There is not much on Pattern Recognition in the
conferece. In IP, there is none. (Bolle) We tried to get Pattern
Recognition papers, but the ones submitted were bad.
Discussion: not wanting quotas, but the reviews were done by
qualified people. Was it the history of the meeting (rejecting past
PR papers) or that no good ones were there. Question is whether we
want Pattern Recognition papers or not. Image Processing has a new conference. Issue is whether
PR authors believe that their papers will be treated properly.
Q: Definition of Pattern Recognition -- at ICPR the Pattern Recognition subconference. Desire to have
some emphasis on the fact that Pattern Recognition papers are acceptable. Need to
indicate more specific topics within Pattern Recognition (High dimensional
classification, stocastic, ...). One role of prog. comm. is to be
broad and to draw in the papers, but good papers must be submitted.
CVPR has changed its flavor, so that many do not send papers here
because they don't fell that they will get a fair hearing (some
cynicism that rejection means the paper is good). Issue of
limitations of the number of allowable papers (3 and 4 tracks...).
Q: did it not include good people (e.g. document image analysis --
no one who is strong). Q: regarding who was on the ProgCom -- not
known until late in the process.
(LShapiro) Comments on next progComm.
() Deal with the number of papers, reason for multiple tracks.
(Freeman) Cultural problem, once a conference doesn't have some
papers, the papers and attendees are not there. It takes more to
get them to become active again. Maybe it needs a separate
conference or separate track to make such papers be submitted.
(Nevatia) The basic issue is one of size -- PR/CV/IP -- and a
question of whether one conference can handle all of them is another
(Haralick) <b>Motion</b>: Change the name of the conference to Computer
Vision. (i.e. drop Pattern Recognition from the name). Seconded (Shaprio).
<b>Vote</b>: Failed (voice vote)
<b>Motion</b>: to have a separate track CVPR to encourage submission of
papers that have been excluded in the past (PR, IP, applications).
Discussion: no way to have third track in the hotel in the Seattle
meeting. But there is a 1/2 track for "experimental" papers/tracks.
Not giving the process enough time, and should wait. Three
dimensions are being confused (selectivity, topic area,
theory/applied). Are separate tracks separate budgets, proceedings,
etc. --- Current motion is only to broaden the meeting. Separate
conferences is a terrible idea -- applications drive CV. More
discussion on what it means to broaden the meeting, but are the
papers as good. In this one Doc Anal had 2 submitted and 1 is in.
Should we legislate the sessions.
Amendment: Accepted the friendly Amendment to make it apply to the
next possible meeting with the 94ProgComm making sincere effort.
<b>Vote</b>: Voice Vote: Passed.
Tanimoto comment, there are many things that can be done and are
Freeman comment there is not a CVPR95.
#3. Governance issues: Proposal for 12 member board (Haralick)
Executive Board for decision making. History, there was an
executive board (10-12) that would meet and make decisions.
(Haralick) as chair invited everyone to attend and vote, etc. to
open up the process. But we are not keeping track of who comes and
who does not. Problem of membership, who is a member, who is not?
Who are members for IAPR issues. We do not have a good structure of
who is a member and who can vote. IEEE does not regard this as a
meeting of the PAMI TC membership when compared to the PAMI journal
publication. So an issue of who are the members so that business
can be done between meetings. There is a need for other meetings.
Some suggestion (from IEEE-CS) to have an executive committee and
votes among the membership by mail and not just who comes to the
conference. An issue is who is a member? and a structure of a
group to make decisions and bring them to the membership.
Establish an executive committee of 12 appointed by the
chair of the TC, then the EC can bring the membership issue to the
current entire membership and to establish how the 12 would then be
continued. With possibly different alternatives for the future of
the EC and TC.
Discussion of who is a member of the TC (according to IEEE) the IEEE
list is a mess. Q: (Kim Boyer) not a formal definition, or not an
enforcement of the existing rules. A: No formal way to get on the
list and By-Laws are not being followed.
Q: Motion to attempt to fix something, what is being fixed?
A: When crucial decisions need to be made must be made by the
Discussion: General membership does not know what a TC is. There is
a lack of awareness of what the TC membership is.
Physical mailing list (TC, Transactions). Issues of cost.
Current motion is an EC, which is a good thing that is willing to
fight the issue with IEEE. IEEE is not forcing structure on us,
IEEE is recommending that all TCs have some structure that will make
(Grimson) We do have an EC as the 4 officers (C, VC, S, T).
<b>Move to Amend</b> to ask the current 4 members of the TC to go to
IEEE to resolve this.
(Not friendly, wants to have it broader).
Discussion on who they would report to?
<b>Vote on Amended motion</b>: <b>Carried</b>.
(Tom Huang, Eric Grimson, Avi Kak, Keith Price)
Comment that the TC list should be updated and maintained.
#4. Report ICCV 1993 (Huang) (Faugeras) In Berlin, Attended by
250-280. The only problem was the representation of US universities
an research centers. Deemed to be a success. There was a meeting
of the "Pami TC" at the conference and the next ICCV conference
between San Francisco, Rockies, Pittsburgh, and Boston. Next ICCV
in Boston in June 1995.
#5. Report on 1992 Applications of Computer Vision (Bhanu). First WS,
70 people, desire for the meeting to continue. 90 submissions, 2 on
Doc anal, 1 on medical. Panel, single track.
#6. Status on CVPR94 (Shapiro, Bowyer). Seattle, June 21-23, 4
tutorials (Monday), 3 Workshop planned (Visual Behaviors; Role of Function
in Object Recognition; Multi-Dim medical analysis(F)) Still space
for Saturday. Publicize the names of the program committee!
Comment that posters had 2 pages -- that was low. Possible same
number of pages. Find a way to recycle the money to the TC by
charging for extra pages (money goes to program committee). Discuss
the proceedings on CD.
#7. Upcoming CAD-Based vision (Ikeuchi). 1994, February 8-10 (W-F)
Seven Springs Ski Resort.
Morning session, afternoon Ski, evening session. July 1 deadline.
#8. PAMI and IPAR Newsletters (Rosenfeld and Kasturi)
(Rosenfeld) Newsletter was discussed earlier, crazy mailing list
(TC) and good mailing list (PAMI). IEEE-CS charges about $2.00 per
newsletter. When electronic was offered, only 100 responded. The
TC is responsible to mail out the IPAR newsletter. Currently twice
a year with the IAPR being the bulk of the newsletter. For a while
it when to the 10K list, but that is expensive.
Proposed to create a short list of US members that attend CVPR,
ICCV, and ICPR (this year) to get the US membership for the IAPR
(Need an immediate solution.)
Discussion as to whether this is the list for the IAPR "society"
Amendment to start as electronic and keep in this form.
(Human vision list has 1K members.)
Clarification that IAPR newsletter must be sent to the US community.
Problem is not the mailing list, but the problem is that we don't
get the proper arrangement with IEEE. Fundamental issue is that we
do not have dues, either dues for the TC or the Robotics with an
extra $5 for dues.
(Grimson:) PAMI generates 300K for the CS, but does not get any
budget to send out a newsletter.
Discussion of making the move to become a society, we must move to
become a society (the officers).
US is a category C, (needs 800 members). With smaller number dues
decrease and representatives.
The US needs a separate Society that has the power.
<b>Motion</b>: <b>Shafer</b>: That we resolve that the EC of the PAMI TC
work with IEEE to work on a better arrangement to have greater
control over its income, its mailing list, its activities (including
the journal), such as a society or whatever.
Can we take the journal? Threaten to pull out and see what happens.
Is it strong enough? Talk to those involved in Robotics&Auto.
gather the information.
<b>Motion</b>: (Kak) That the PAMI newsletter to those that attend
CVPR, etc., publish in the journal over the past 4 years.
Amendment that email list be created and a separate non-email.
#9. PAMI Transactions (Anil Jain) 1 submission per day, working
hard to get reviews to authors. 80% first review in 6 months. 20%
delayed (due to audience members, etc.). Complain to Anil if there
are problems. Please to review in a timely manner. Authors should
do 3 reviews for all submissions. PAMI subscriptions have decreased
the least (all are down). Down 3%. Trans-C down 10% Possibility
in 5 years that all will be available in CDRom, a trend to CDRom.
Comment on earlier motion, that CS will not like the idea that PAMI
leaves. (Get 35% of profit of the conference only.) (AJ:) It is
possible to separate. PAMI is sponsored by several societies, but
CS gets to net profit (300K). ($24 for members, Libraries 900 of
#10. 1994 Workshop on Multi-Dimensional Biomedical Image Analysis
(Goldgof, Acharya) Friday after CVPR. Q: on name multiple sensors
or multiple dimensional sensors (yes). Should it be multi-modal?
#11. Workshop on Performance vs Methodology in CV. (Meer and Haralick) --
Performance vs. robustness. Applied statistics and others.
Performance characterization issues, analyze problem of numerical
sensitivity. This would be the third in a series (Seattle in 1990,
1991 in Bonn). Saturday before or after CVPR.
#12. Proposal for a second Workshop on Applications of Computer Vision (Dyer).
On the MTW after Thanksgiving in 1994 (November 28-30) (Hanson as
general chair/ Burt and Flinchbaugh as program chairs).
#13. Proposal for a 1995 Workshop on CV/Motion Analysis (Medioni,
Negahdaripur, Goldgof). 1995, there is no CVPR, no ICPR, ... in the
past there has been a Workshop on Computer Vision as an outlet for
other good papers.
There is a Workshop on Motion every about 3 years. Discussion of Florida
(Miami). (Tom Huang, General Chair). Two tracks, one on Motion on
(Possible dates, in October 1995; or Week before Thanksgiving.)
#14. CVPR96 Proposal (Bhanu). Proposal to have CVPR 1996 in
California (San Francisco) (return from 1995). Dyer/Ikeuchi as
ProgChair, Bhanu as General Chair.
Comment that no Pattern Recognition in 1995.
<b>Motion</b>: That the 1995 Workshop expand the scope to include it in the
<b>Vote</b>: Yes: 21 No: 20.
#15. Announcement of "Physics-Based Vision Newsletter" (Shafer)
A gang of 7 is assembling an all-electronic on physics based vision.
Open question is whether there is running commentary (currently no).
Shafer, Wolff, Healy, Nayar, Woodam, Tomanaga, xx. Initially
maintain an email list. Send mail to Glen Healy.
On CVPR94, the area of Image Processing is not called out.
<b>Motion:</b> (Haralick) Specifically call out Image Processing and its subtopics
in the next Call for papers (94).
Duplicate submissions -- policy is stated on the call for papers for
<b>Motion</b>: That the program committee reconsider that the Prog Comm
reconsider the policy on duplicate submissions. (Keeping the
prohibition on dual publications.)
Discussion that people will break this policy and to make rules
that. The number that attend both is small. Comment about the
decision last year. Just because people want it doesn't mean it is
good -- this leads to multiple journals, etc. Strive for quality.
#16. Election of a new PAMI-TC chair (Huang).
Jake Aggarwal is now President of IAPR.
Again for Terry Boult for local arrangements.
The term is over for Avi Kak. The nomination Committee of Tom
Huang, Jake Aggarwal, A Rosenfeld, Avi Kak, Herb Freeman. Nominated
Linda Shapiro. No. Nominations from the floor.
<b>Carries</b> by acclamation.