To analyse, apply and illustrate the directions of exploration carrying respectively the individual, common and collaborative representations with dimensions, to the transfer of roles and their sharing between humans systems and the other, it seemed to us that the activity of a dancer, in a setting in a scene at the time of an event could be used by us as a support of exploration and an illustrative example.
This scene setting is accompanied and reinforced using physical devices of observation and capture of information on the one hand (communicating purposes), and of production of proposals and mediations on the other ( intelligent agents ).
The perception of the spectacle relates to analysis by the system of all that is external to it, i.e. in particular what occurs in the scene and in public. That relates to, in particular, all displacements of the dancer, his body and gestural attitudes, his physiology (rate of heartbeat, breathing , surface tension of the skin, etc), words, and more generally all his interactions with the environment (devices, other dancers, public, etc).
This analysis is a stage of abstraction representing rough perceptions of the system in terms of qualities as defined in the preceding paragraph, then in partitions of qualities. It will be carried out by considering each medium of expression as the support of a communication whose preliminary phase of analysis will make it possible to define the words, syntax and semantics (Figure 23.2).
There are various techniques of capture of movement but their performances are not equivalent when taking into account our criteria:
perception : must be as precise as possible but it is possible to work and do interesting things even with weak precision;
response time : the interaction must be made in real time; perception must thus be not only carried out in real time but also leave calculating time for the other operations necessary ( decision, realisation of the performance );
the device : must be most discrete possible so as not to centre scenography around technology, and must be as transparent as possible so as not to block the dancer in his choreography.
Within sight of this synthetic presentation of the various existing solutions, our choice was made, as a first approach, of the least constraining technique for the most economic dancer, and analysis of video images. So thereafter, if the precision obtained proved to be insufficient, other heavier techniques could be considered (Figure 23.4).
In the device considered, a camera (Webcam type or more sophisticated) films the scene in a continuous way. The exact position of the camera (face to face with the scene, at 45 degrees as in the figure above or with the balance of the scene) will be specified according to results' of the first experiments. A 3D model of the scene will have to be developed to facilitate the recognition of the position and postures of the dancer. Recognition of the movements of the dancer is carried out in several stages:
all primitive descriptors are extracted from the video image, including box, direction of transfer, (high/low, ahead/back, left/right), dynamics (acceleration/deceleration). These descriptors can be supplemented by information of reliability. Webcams of the market are sold with software which already does part of this work;
these descriptors can then be exploited, complementing structural information on the human body, for the determination of the position of the dancer in the scene and the recognition of postures.