PROJECT PROPOSAL

Bibliography

@Article{Badler/Smoliar:1979,
  author=       { Badler, Norman I. and Smoliar, Stephen W.},
  year=         { 1979},
  institution=  { U. of Pennsylvania 2-->National U. of Singapore},
  title=        { Digital Representations of Human Movement},
  journal=      { ACM Computing Surveys},
  month=        { march},
  volume=       { 11},
  email=        { badler@central.cis.upenn.edu; Smoliar@ISS.nus.sg},
  annote=       {
The techniques of representing a human being as a computer generated 
graphic entity are discussed. A brief description of the Labanotation 
used to presicely quantify body postures is given. It is emphasised 
that although direct representation in a 2D space is possible, in the 
general case, the better approach is to construct a 3D model and then 
project it to a plane. Three aspects are discussed:
1 Representations of the human body
   this is done by the following methods:
    stick figures : the simplest and most unimpressive
    surface models: excellent results, but slightly unrealistic and 
                    a fair share of blemishes.
    vomule models : the human body is decomposed into primitive solids
                    such as cylinders, spheres, and ellipsoids.
2 Representation of movement
  Given a model for a human body, the process of animating it involves 
  producing a succession of frames, each slightly different from the 
  previous. This is attained by key frames, where a set of important 
  frames is provided and the intermediate parts are interpolated; by 
  movement functions where labanotation is used to choreagraph the 
  motion; and simulation where the mechanics of the human body are also
  encapsulated in the model.

3 Finally, an architecture for such a system is discussed. 
-k.anandsudhakar feb/2k }

}

@Misc{Geib/Levison/Moore:1994,
  author=       { Geib, Christopher and Levison, Libby and
                  Moore, Michael B.},
  year=         { 1994},
  institution=  { upenn},
  title=        { SodaJack: An Architecture for Agents that Search for 
                  and Manipulate Objects},
  month=        { january},
  email=        { (geib,libby,mmoore)@linc.cis.upenn.edu},
  annnote=      {			
This paper deals with the problem of an agent whose aim is to search 
and undertake manipulation tasks in an environment. They have 
implemented the approach in a system called SODAJACK, which does the 
animation. The agent receives as input high level commands like 
"fetch the scoop" and the system has to figure out the exact low-level 
action to do the job. This involves a knowledge about the possible 
locations of the scoop, plan a route to those, explore them, and then 
finally the act of lifting it up. 
The system has been divided into a hiearchy of three planners that 
respond to the input goal, and give as output the action outline to 
acheive the goal. The task division is like this:
1.search planner, it converts the goals into a plan to search.
2.object specific planner, this relates to each search plan by the 
search planner a particular object and undertakes feasibility tests 
for the action plans generated.
3.hierarchical planner(ItPlans), this supervises the other two and 
delegates the control first to search to get a plan and then to the 
object to make it specific.  -Vikrant Kumar 11/02/2000 }

}

@Misc{Salesin/et.al:1994,
  author=       { Christianson, David B. and Anderson, Sean E. and
                  He, Li-wei and Salesin, David H. and 
                  Weld, Daniel S. and Cohen, Michael F.},
  year=         { 1994},
  instituiton=  { 4U. of Washington; Microsoft Research, Redmond
                  2-->Stanford},
  title=        { Declarative Camera Control for Automatic
                  Cinematography},
  email=        { (dbc1,lhe,salesin,weld)@cs.washington.edu;
                  seander@stanford.edu; mcohen@microsoft.com},
  annote=       {
For long programmers haven't made use of cinematographic principles in 
computer animations. The authors try to fill this gap by making the 
rules of cinematic storytelling lend themselves easily to programming. 
For this purpose they have formalized the rules into a  Declarative 
Camera Control Language(DCCL). Such a thing will be very useful as it 
will allow programs to present a dramatic point of view aesthetically.
The authors first introduce the language of the cinema like the breakup 
of a film into scenes and shots, shots being the smallest unit. Another 
thing is the placement of the camera, which depending on the scene can 
be apex, internal, external or parallel. Cinematograohers have 
identified certain field of views of the shots which give pleasing 
results. And then there are certain constraints on a shot which should 
be satisfied like parallel editing and break movement. Next comes the 
concept of idioms, which is the way cinematographers describe 
situations in a film. DCCL is an attempt to formalize this idiom. 
The DCCL is composed of four basic components fragments, views, 
placements and movement endpoints.Fragment is the time interval during 
which the camera performs a simple motion. A simple shot may comprise 
of one or more fragments.

Next they define the Camera Placement System(CPS). The CPS is a three 
stage pipeline consisting of 
1.the sequence planner
2.the  compiler
3.the heuristic evaluator.
The basic aim of the CPS is to give the camera positions depending on 
the input of the positions of the various interacting entities. And 
the authors have implemented this approach in a video game.
The authors succeed to bring to the highlight the importance of using 
the cinematographic techniques in computer animations so as to make the 
experience more enriching.             -Vikrant Kumar  11/02/2000 }

}

@Article{Lozano/Lozano-Perez:1996,
  author=       { Lozano, Oded Maron Tomas and Lozano-Perez, Tomas},
  year=         { 1996},
  institution=  { 2mit},
  title=        { Visible Decomposition: Real Time Path Planning in 
                  Large Planar Environments},
  journal=      { AI Memo},
  month=        { Januaray},
  www=          { ftp://ftp.ai.mit.edu/pub/users/oded\
                  /papers/planning.ps.Z},
  email=        {oded@ai.mit.edu,tlp@ai.mit.edu},

  annote=       {
This paper deals with the use of visibility graphs to do motion 
planning.
	  -Rajesh Rajasekar 2/2000}
}

@Article{Hoffman/Hopcroft:1987
  author=       { Hoffman, Christoph M. and Hopcroft, John E.},
  year=         { 1987},
  institution=  { Purdue-cs; Cornell-cs},
  title=        { Simulation of Physical Systems from Geometric Models},
  journal=      { IEEE J. of Robotics and Automation},
  month=        { june},
  vol=          { RA-3},
  annote=       {
The mechanics of simulation are discusssed. -k. anandsudhakar feb/2k}

}