1. STATEMENT OF PURPOSEThe purpose of AVISA, the International Speech Communication Association (ISCA) Special Interest Group (SIG) on Audio-Visual Speech, shall be to promote interest in audio-visual speech; to provide members of ISCA who have a special interest in audio-visual speech with a means of exchanging news of recent research developments and other matters of interest in audio-visual speech; to create a scientific link between researchers and others working in the various disciplines associated with audio-visual speech processing by humans or by machines; to sponsor meetings and workshops in audio-visual speech that appear to be timely and worthwhile, operating within the framework of ISCA's by-laws for SIGs; to facilitate better understanding of the cognitive processes involved in the production and perception of audio-visual speech communication through the audible and visible gestures of the vocal tract and of associated body gestures; to facilitate the modeling of speech gestures for realistic animation of synthetic actors and for automatic recognition of linguistic and communicative gestures transmitted through the auditory and visual modalities; to share linguistic knowledge through the study of audio-visual speech in as many languages as possible, including sign languages and cued speech; and to provide and make available resources relevant to audio-visual speech, including audio-visual data corpora, physiological data corpora, analysis tools, analysis and generation software, video tapes, research papers and generated data.
2. BACKGROUNDIn the last several years there has been considerable excitement and energy in the area of audio-visual speech within the speech community and beyond. This is due in part to the general interest in multimodality in the area of cognitive sciences, the progress made in modeling the physiology of speech production and facial structure and motion, and also to the new possibilities created by multimedia computers that permit the design of new investigative tools (e.g., microcameras for easy video capture of lip gestures, synthetic talking heads, etc.). In addition the design of human-machine interfaces and virtual humans and the continued interest in the motion picture/animation community have the potential for providing a promising market for the development of tools, techniques, and concepts. Audio-visual speech is intrinsically a pluridisciplinary area of inquiry. As such, its thorough study requires extensive collaboration across various disciplines such as phonetics, psychology, psycholinguistics, linguistics, pragmatics, physiology, neurophysiology, speech pathology, logopedics, medical imaging, engineering, signal and image processing, artificial intelligence, computer vision, and computer graphics. A most efficient and stimulating way to facilitate such interdisciplinary communication is to organize international workshops where ongoing research is publicly presented. The emergence of an audio-visual speech research community was first observed in Bonas (France) in 1995 where David Stork (Ricoh California Research Center), with the help of Michael Brooke (Univ. of Bath, UK) and Christian Benoît (CNRS - Univ. Grenoble, France) as co-directors, organized a two-week NATO-ASI conference on "Speechreading by Humans and Machines." Scientists representing academia and industry presented, discussed and, for the first time, realized the breadth and richness of the audiovisual speech processing domain. Topics included speechreading and bimodal speech perception by normal and hearing impaired listeners from psychological, neurophysiological, phonetic, and didactic perspectives, together with work on automatic speechreading and audio-visual recognition by machines in signal and image processing, biometrics, computer vision, and sensory fusion perspectives. In 1996, Lynne Bernstein (HEI, Los Angeles, USA) and Christian Benoît organized a special session on "the multiple senses of speech perception" at the International Conference on Spoken Language Processing in Philadelphia. In September 1997, a third meeting, the "First ETRW on Audio-Visual Speech Processing", was organized by Christian Benoît and Ruth Campbell (UCL, London) in Rhodes (Greece) as a satellite of the 5th Eurospeech Conference. Finally, the second AVSP, organized by Denis Burnham (UNSW, Australia), Jordi Robert-Ribes (CSIRO-Sydney, Australia) and Eric Vatikiotis-Bateson (ATR-HIP, Kyoto, Japan) will take place near Sydney (Australia) over three days in December 1998, as a satellite of the 5th ICSLP Conference. In addition to providing venues for the presentation and discussion of ideas and research, these meetings have resulted in published proceedings and follow-up publications. These are but a few examples of the impact the burgeoning interest in audio-visual speech processing by humans and by machines is having around the world. It is obvious that a coordination of efforts is needed to organize regular events, promote them, and keep this growing audio-visual speech community informed of relevant activities and various issues.
3. ELECTED OFFICERSThe elected officers of the SIG shall consist of a Chairperson, a Vice-Chairperson, and a Secretary. The Chairperson, Vice-Chairperson and Secretary shall be members in good standing of ISCA. The term of all elected officers of the SIG shall be two years. They can be re-elected, but consecutive appointments are limited to four successive terms.
The duties of the Chairperson shall be:
- to have primary executive authority over actions and activities of the SIG;
- to prepare an annual written report on the activities of the SIG for the ISCA Board before July 1st each year;
- to designate a Liaison Representative for the SIG.
- to appoint an Advisory Committee of at least five (5) other SIG members.
The duties of the Vice-Chairperson shall be:
- to serve in the role of the Chairperson when the Chairperson is unable to attend to actions and activities of the SIG;
- to assist the Chairperson with SIG business.
The duties of the Secretary shall be:
- to maintain a membership list of the SIG;
- to be responsible for any funds awarded to the SIG by the ISCA Board;
- to collect and manage any dues that may be required by the organization;
- to be responsible for communicating with members of the SIG;
- to be responsible for answering inquiries about the SIG;
- to present a written annual report on the SIG finances to the ISCA Board.
The Liaison Representative shall be primarily responsible for communication with the ISCA Board.
The Advisory Committee will provide guidance to the SIG officers on matters of SIG business and to issues related to audio-visual speech.
4. ELECTION OF OFFICERSAll officers of the SIG shall be elected by a vote of the membership. The vote shall take place at least 3 months before the expiration of the terms of the officers to be elected, with at least 4 weeks notice of nominations to SIG members. In the event of a casual vacancy, an extraordinary election shall be held for an officer to complete the current term of office. Candidates must be nominated by two members, must accept nomination, and must be members in good standing of ISCA. Votes may be entered by electronic or physical mail, and votes arriving by the notified closing date will be counted by a returning officer appointed by the Chairperson and agreed to by all candidates. Nominations shall be called at least three months prior to the election, and shall close two months prior to the election. All members shall be notified of the election date and the nominations at least four weeks prior to the close of voting, and results shall be notified to the members within four weeks of the close of voting. Notification shall signify dispatch of mail by the medium and to the address last notified to the secretary.
5. DUESAt present, AVISA will not be collecting dues. The decision to collect dues for this SIG will be left up to the membership of the SIG, subject to the approval of the ISCA Board.
6. REFERENDAA change to this constitution, or a motion on any other matter, may be initiated by five members (petitioners) in writing or by agreement of the Chairperson, and Secretary (on the suggestion of a member or otherwise). The motions will be put to the members within six months of receipt of a petition, if the petitioners so request. Otherwise, constitutional amendments and other motions will be put to the vote in conjunction with the regular elections. Members will be given notification at least three months in advance of any referendum, and further motions or amendments proposed by two members shall be accepted up to two months before the close of voting. Votes may be entered by electronic or physical mail, and votes arriving by the notified closing date will be counted by a returning officer appointed by the Chairperson and agreed to by all petitioners. All members shall be notified of the date of the referendum and the proposed motions and amendments at least four weeks prior to the close of voting, and results shall be notified to the members within four weeks of the close of voting. Any change to this constitution requires a two-thirds majority of those eligible to vote and is subject to approval by the ISCA Board. Any other matter requires a simple majority of those actually voting. Notification shall signify dispatch of mail by the medium and to the address last notified to the secretary.
7. RESOURCE MANAGEMENTIn providing and making available any resources as defined in Clause I, and especially any held and distributed electronically, the SIG shall follow proper procedures with respect to matters like copyright, registration and data protection as is appropriate to each case. Furthermore, where the SIG may be deemed the proprietor of a resource, as the SIG is a subordinate body of ISCA it is understood that ISCA is ultimate proprietor and therefore other members of ISCA who are not members of the SIG may have access to the SIG's resources and the ISCA Board may itself make available such resources even to non-members. The SIG shall not enter into any formal or potentially legally binding contracts without the approval of the ISCA Board (or the ISCA Treasurer acting under delegated powers). The SIG shall not enter into any financial commitments for such resources without ensuring that the moneys required are available.
8. DEDICATIONThe AVISA SIG was originally conceived by Christian Benoît and is dedicated to his memory.
Created by Philip Rubin on September 17, 1998
(revised on March 8, 2000)