Crowd Sensing

The project proposes a dialog between a group of people and the machine, understanding the group as an organism, as a single intelligence.

The interface selected to stablish this dialog was a visual projection that reacted to the input fed by video recognition. The first approach to the installation was to set the two main elements (projector and camera) in the same place above the crowd to achieve the desired perspective for the input and the output. for this purpose the team developed an artifact that could act as a holder for this two components.

The artifact had to be quick to assemble using the available materials and machinery and hold enough weight in the simplest possible way since it was gonna suspend a heavy object above people's heads.

The previous set of images are a simple assembly guide for the mentioned object. The projector holder was developed to be assembled without any screws or adhesive.

The holder is produced using 15 mm thick plywood. Tolerances for the cutting file should be adjusted prior to CNC milling, based on actual material thickness.

The object is intended to be hanged and secured from any kind of structure. For this purpose the artifact has 3 holes intended to make it easier to level the projector as desired regardless of the mounting spot as well as the shape itself of the piece has extra lateral space from which is possible to tie any extra rope.

After testing the artifacts capability to hold a projector from above in a secure manner it was possible to begin crowd recognition trials.

For the first trials, it was decided to use a cellphone camara to make a wireless connection to the computer where the code was being developed. This allowed a quick trail and error work flow.

After a few iterations, the software was tuned to be able to recognize only human bodies and exclude surrounding objects.

The next step was to program a responsive stimulus delivered by the machine in order to interact with the crowd and stablish the desired dialog

For this purpose, the bounding boxes identifying human bodies were replaced by colored circles that engage with each other based on proximity.

The idea of this way of interacting is to test the possibility of influencing in a crowds behavior using visual signals.