Subject: FW: Position paper (txt version)
Date: Thu, 10 Feb 2000 16:40:52 -0000
From: D.R.Morse@open.ac.uk
To: anind@cc.gatech.edu

Anind,
    I'm sorry I forgot to send this to you....

Best wishes,
David.

-----Original Message-----
From: Edson Moreira [mailto:edson@icmc.sc.usp.br]
Sent: 04 February 2000 22:34
To: David Morse
Subject: Position paper (txt version)

David,
Just in case you had problems with my position paper.
Regards,
Edson
--------------------------------------------------

Embeded Video Content and Context Awareness

Edson dos Santos Moreira
Universidade de São Paulo
ICMC-USP, Cx Postal 668
13566-570 São Carlos SP
Brasil

Introduction

In modern computing, the user interacts with the computer through the
control
 of tasks and the environment in which those tasks are inserted.
Communication,
dialogue, perception and cognition  must be considered on evaluating the

interaction [Preece, 1994; Pressman, 1995].

Traditionally, events and relationships are analysed and modelled to
build interfaces.
As computing is used ubiquitously [Abowd, 1999], the design of
interfaces changes
to demand further capabilities both from the designer and the
infrastructure available.

This position paper speculates on the use of content-based distributed
multimedia
protocols features for the enhancement of human-computer interaction,
providing means
to improve interactivity. We consider this issue important for the
workshop as the
boundaries between the network and the applications which are suported
by it get more
and more complex.

Embeded Video Content

Adding meta information about content to audio and video streams results
in versatility
for the application designer. Three facts support this statement:

· The integration of several medias is important in the process of
development of
hypermedia applications. For instance, in many application domains (e.g.
in education,
business and entertainment), the simultaneous presentation of multiple
flow of data are
expected and flexible (high level) means of synchronization are sougth;
· The interaction of the user with the media objects being presented (be
they natural
or synthetic) is seen as a bonus and a important requirement for the
success of an application;
· The above characteristics influence the whole process of coding,
storing, distribution
and presentation of medias.

MPEG-4 and Related Technologies

MPEG-4 [ISO, 1997c] has been designed to provide the description and
management of media,
as well as well as the control of synchronisation and retrieval of
medias contained in a
stream being transmitted. Other features like the interactivity
management and the description
of scenes in terms of temporal and spacial behavior are also
incorporated in the protocol.

In a stream, it is possible to, randomly, search  for video sequences
and its contents.
Aspects like compression techniques, transparency coding and colour
standards and resilience
to errors have been specified. Through the specification of profiles,
tools can be implemented
for the components of the standard. Some of these profiles:
· Visual Profiles (eg. Basic animated 2D)
· Audio Profiles (eg. Low Rate Sybthesis)
· Graphics Profiles (eg. 2D Graphics)
· Scene Description (eg. VRML Scene)

MPEG-4 also provides an API for the access to the media objects
contained in the streams.
It is Java-based and is called MPEG-J. Using this interface, applets can
travel with the
streams and can be used by the user (or user environment) to adapt the
presentation of that
particular information, possibly with inputs which will depend on the
context. XML
[Johnson, 1999] will certainly be very important in developing
applications using these
features of MPEG-4.

Open Questions

MPEG-4 is a technology under development. We envisage that its use can
be useful to
potentialise the infrastructure in order that context awareness be taken
in consideration in
the design of distributed multimedia applications. The questions that we
put forward to
discussion (supposing that some of the initial "W" questions have been
answered):
· Which kind of context can be properly mapped into meta-information in
media objects data
stream;
· How this approach compares with possible alternatives;
· What is the overhead imposed on the system (coding, transmiting,
decoding)? Does it pay off?
· How to structure the meta-information in order to ease the access in a
generic way; MPEG-7
[ISO, 1997a] is a possibility.

References

(Abowd, 1999) Abowd, G.D.; Software Engineering Issues for Ubiquitous
Computing. Proceedings
  of International Conference on Software Engineering (ICSE'99). May,
pp. 75-84.
(ISO, 1997a) ISO/IEC JTC1/SC29/WG11 N1921, Third Draft of MPEG-7
Requirements, 1997.
  http://drogo.cselt.stet.it/mpeg/public/w1921.htm
(ISO, 1997c) ISO/IEC JTC1/SC29/WG11 N1909, Overview of the MPEG-4
Version 1 Standard, 1997.
  http://drogo.cselt.stet.it/mpeg/public/w1909.htm
(Johnson, 1999) Johnson, M.: XML FOR THE ABSOLUTE BEGINNERS, 1999.
  http://www.javaworld.com/javaworld/jw-04-1999/jw-04-xml-p.html
(Preece, 1994) Preece, J. et al.: HUMAN-COMPUTER INTERACTION.
Addisson-Wesley, 1994.
(Pressman, 1995) Pressman, R. S.: ENGENHARIA DE SOFTWARE. Makron Books,
1995.