Minutes of the Multimedia Signal Processing Technical Committee Meeting ICASSP 2009, Taipei, Taiwan Wednesday, April 22, 2009   Attendees: John Apostolopoulos, Oscar Au, Andrea Cavallaro, Yen-Kuang Chen, Philip A. Chou, Francesco G. B. De Natale, Jiebo Liu, Petros Maragos, Marco Tagliasacchi, Wenjun Zheng (10, quorum not achieved) Guests: Dinei Florencio (MMSP'09 Co-General Chair), Mercy Kowalczyk (SPS Executive Director), Lisa Schwarzbek   Phil Chou called the meeting to order at 11:40 am.   1. INTRODUCTIONS 2. REVIEW AGENDA 3. REPORT OF THE CHAIR 3.1 TC Report to SPS [Phil, for Anthony] Every TC reports to the SPS Technical Directions Board, chaired by VP for Technical Directions, Alex Acero. Every three years, the report is a major report subject to formal review. MMSP's review is coming up next year. This year was a minor report, reporting on: * Membership: 8 new members * Distinguished Lecturers: none nominated * Awards: several nominated, only Education Award (Bede Liu) successful * Conferences and Workshops: MMSP'08, ICASSP'09, MMSP'09 budget (expenses, surplus), papers (submitted, accepted) * Publication support: MMSP TC members on T-MM, T-IFS, SPM, SPL editorial boards * Website * Other activities 3.2 Subcommittees re-established [Phil, for Anthony]   New subcommittee chairs / members: * Membership: Y.-K. Chen / A. Cavallaro, T. Ebrahimi, J. Luo, K. Rose * Conference: M. Wu / O. Au, C.-W. Lin, C.-Y. Lin, F. Pereira, D. O. Wu * Awards: A. Said / P. Maragos, B. Pesquet-Popescu, P. A. Chou * Technical Directions: E. Steinbach / J. Bloom, F. De Natale, M. Tagliasacchi The TC will also be looking for various liaisons. 3.3 Web site updated [Phil, for Anthony] The TC website (http://www.signalprocessingsociety.org/technical-committees/list/mmsp-tc/) has been updated to include new members, new subcommittees, MMSP'08 minutes, new publication support, and revised EDICS. 4. WORKSHOPS 4.1 MMSP 2008 (see TC meeting minutes from Cairns) [Phil, for Jian] Submitted papers 325 (up from 165 in 2007). Acceptance rate 55% (higher than past years). Registered attendees 198. Sponsorship for best paper awards from Microsoft, Cannon, NICTA. Budget US$70,102. Expenses to date US$61,440. Anticipated surplus US$8,662 (14 %). Books are almost closed; just need to transfer surplus to IEEE. This years workshop ran two tracks for first two days, one oral and one poster, then all oral on last day. Somewhat different from previous years that had dedicated poster slots running all days. Suggest minimal overlap between poster and oral going forward and running posters every day. 4.2 MMSP 2009: update on technical program and other plans [Dinei] Keynote and overview speakers have confirmed (Fred Juang on acoustics, Inald Lagendijk on secure signal processing, Joe Paradiso on media sensors, Vivek Goyal on compressed sensing, Anthony Vetro on 3D multimedia, Liliana Nakonechnyj on Brazilian TV). Special sessions on telemedicine, compressed sensing. Panel on compressed sensing. Top 10% Award instituted. (There was a 40-50% increase in visits to the web site after announcing the Top 10% Award.) Number of visits to web site is high: in March, 1400 total visitors, 1300 new visitors. Original submission deadline was last Friday. Current deadline is May 1. 62 submissions so far. Hard to predict final number of submissions. 4.3 MMSP 2010: submitting budget, securing venue [Beatrice] Venue secured. Currently contracting with venue: investigated with INRIA and SPS the kind of contract, and looking at the financial details. Preparing call for papers. 5. SUBCOMMITTEES 5.1 Awards [Phil for Amir] Amir Said is the new chair of the Awards Subcommittee, with members Beatrice Pesquet-Popescu and Petros Maragos. Phil is advisor during transition. * Signal Processing Magazine Best Paper Award Nominated “An Introduction to Compressive Sampling,” by Candès and Wakin. Received: No. [Received by Koetter, Singer, and Tuchler, "Turbo Equalization"] * Technical Achievement Award Nominated Aggelos K. Katsaggelos. Received: No. [Received by Van Trees, Unser] * Education Award Nominated Bede Liu. Received: Yes. * ICASSP 2009 Student Paper Award Nominated Rate Efficient Remote Video File Synchronization. Finalist but not winner. Nominated Phase Diffusion for the Synchronization of Heterogeneous Sensor Streams. Received: No. In the past two years, paper award nominations were collected in July and voted upon in August, whereas major award nominations were (offset by one month from paper award nominations) collected in August and voted upon in September. The schedule has been tight with difficulty getting a majority of the TC to participate. This year we may spread the process out from 3 months to 5 months (i.e. May - September), with 2.5 months each for non-overlapping paper and major award nomination processes, at the risk of "award fatigue." The membership (John, Beatrice) felt that a short schedule earlier is better. 5.2 Conference [Oscar for Min] There has been much discussion by email on post-ICME planning. (See Agenda item 6.) 5.3 Membership [Yen-Kuang] The following schedule is anticipated: From now until two weeks before ICME, we want to identify (by email discussion among the TC) what kinds of representation we are looking for. Then from ICME until two weeks before MMSP, we will call for nominations. At MMSP we can have discussion of the candidates. Two weeks after the MMSP meeting, we will have electronic voting like last year. We are targeting 7-8 new members, but that could change depending on post-ICME needs. Jiebo asked if we have data on the areas of expertise of current members. YK said that the Membership subcommittee will send out email to collect updated EDICS areas from TC members and will make this information available. Wenjun stressed that is it important for us to decide on which areas we need to improve representation. YK noted that over time, a skewed distribution in interests tends to magnify itself due to positive feedback. Andrea would like to know the number of ICASSP papers submitted to each of our EDICS areas. Phil said that Anthony should have this information and will follow up. YK noted that first the EDICS need to properly reflect our needs. Oscar noted that it is important to have good people in the less represented areas if we are ever going to start attracting good papers in those areas. 5.4 Technical Directions [Francesco and Marco] The subcommittee started to review the EDICS. Starting point was very good, but some EDICS items are too specific, while some are too general. Paid some attention to watermarking and security, quality of experience, and whether applications should be split out. Will probably come out with a proposal for the next meeting. Beatrice asked that the subcommittee watch out for the overlap of watermarking and security with the IFS TC. Oscar asked whether it was okay if we overlap EDICS with another TC. Petros responded that the MMSP TC should emphasize the combination of modalities in our EDICS; we need to discover what is unique to multimedia. 6. POST-ICME PLANNING Phil summarized the elements that came out of the email discussion of a possible post-ICME conference: its character or personality; its relation to the MMSP Workshop; its name; its time of year; its launch date; its duration and registration policies; and its location. Mercy clarified the reasons for stopping ICME. It was noted that retaining the name was not a good option since it is associated with a weak event and it would be better to start fresh. Andrea asked whether we actually need to start a new conference. Mercy responded by saying that we do not actually have to start a new conference, and focus on the workshops. Oscar asked whether other TC's in SPS may participate in this if MMSP TC decides not to commit. Mercy responded no. Beatrice asked about other societies. Mercy said it might be possible to create an umbrella event, but it is not clear that any other societies will pick it up. Beatrice asked how other societies share responsibility in the TMM. Mercy responded that SPS is the managing partner, but all four societies have members on the EB. Journal is doing quite well, #1, 2, or 3 in Computing (not Engineering) Index. It is healthy. Jiebo drew some comparsions to the ACM conference and journal. Beatrice and Marco noted the low acceptance rate of the conference. Oscar considered whether SPS should move towards a lower acceptance rate and noted that this wouldn't work for him. It was suggested that journals are more important than conferences, which should have higher acceptance rates. There have been suggestions to increase quality, but we need to think. Dinei examined number of citations for ACM papers and ICASSP papers for past 5 years; he did not find that ACM papers were more highly cited. Suggested that a 10% paper award can combine instant recognition rates with higher acceptance rate. Petros noted that good quality papers at other venues is partly due to allowing 6 pages (or 8 pages with additional payment), and partly due to increased competition (lower acceptance rate). Favors support for MMSP Workshop over ICME, perhaps eventually elevating it to a conference. Should focus on SPS strengths. MMSP Workshop might grow. For example, ICIP started from IMDSP. The first ICIP had 400 attendees, but ICIP 2009 had over 2000 submissions. Beatrice commented that it is difficult to increase quality by changing acceptance rate by changing expectations for just one conference. Reduce from 100 papers to 10 is best way to kill it. Francesco said that best way to increase quality is to be attractive, but we have two strong competitors: ICIP and ICASSP. We need to attract multimedia. We need to understand what kind of works we want to attract. Marco expressed support for Petros' idea: don't need a new version of ICME (to collect rejected papers), MMSP may be a better way. Jiebo noted importance of timing. Wenjun commented that MMSP is successful, but there are still problems with large numbers of submissions. We still need to identify a way to get more papers. Can we leverage other societies to get something unique? Oscar provided more information on ICME: it attracts 400 papers/year and is not a bad conference. Asked that we do not underestimate the potential of keeping the name. It has potential. Major advantage is multiple societies, if we can get this TC to work with other TCs. Andrea asked whether there is there a possibilty to keep ICME. Response from Mercy: From recent talks with Anthony, Presidents, Oscar: one opportunity is to create an umbrella event by SPS. The TCs of the 4 societies would each create a track, and would have two days each, over 4 days, two tracks going at any time. This would mean the societies would have more interest in building good programs. If offered a per-day registration fee, and put in A cities, it could attract industry people. All societies have a different style. What societies do best is high-end technical programs. Do not distract ourselves. Encourage all four societies to work. Those programs would be available to all attendees. Should be fewer no-shows, more dynamic. 7. FUTURE MEETINGS June 2009 at ICME, Cancun, Mexico October 2009 at MMSP, Rio de Janeiro, Brazil 8. AOB Oscar asked how many TC meetings are we expected to attend. Phil responded that the expectation is at least one, preferably more. YK advertised the workshop we are organizing at ICME. 9. ADJOURN The meeting adjourned at 1:30 pm.