Creative Catalyst For Pair Choreography Based On Genetic Algorithm. The system uses a kinecct senser to capture initial input, and employs a genetic algorithm to generate novel movement catalysts for the choreographers. This is the client part of the system.
Video Demo https://vimeo.com/manage/videos/193668784
Screenshots
Initialization of Input (Realtime in Kinect Capture)
Generation of Novel Movement Catalyst
Parent Pools and Generation Results in Different Settings
(a) Influence = 0, Intensity = 0 (default setting);
(b) Influence = 100, Intensity = 0;
(c) Influence = 0, Intensity = 100;
(d) Influence = 36, Intensity = 36;
System structure
The whole system includes three layers of structure: Kinect input layer, network connection layer and application layer. The Kinect input layer is the lowest layer. In this layer, we use a Microsoft Kinect 2 motion sensor for the motion capture of the users. This sensor is connected to the socket server in the network connection layer. The socket server runs a C# program which manipulate the Kinect sensor with its SDK. In the network connection layer, there are two different server set up. The web server delivers the necessary files of the website to the client when requested, including a JavaScript file, which sets up WebSocket connection with the socket server. The socket server, which is connected to the Kinect sensor, deals with the connection request, build up the socket connection and deliver the Kinect data. When the socket server is set up, it opens the Kinect server as well as a web socket service for clients to connect. When skeletal data are detected by the Kinect server, the socket server will send the data to the connected clients. So in this way, we build up a real-time connection between Kinect sensor and the browser side. The Kinect input layer and the network connection layer deals with the skeletal data input and transportation. The third layer is the application layer, which means the front end GUI and the logic behind. This layer presents the system to the end-user. It works on the data interpretation and the interface logic control. A genetic algorithm is employed in this layer for the generation of the movement catalyst.
