Arkam is my Bachelor's Degree project. It is based over the Zenith engine for the graphical stuff. The thesis was focused in the programming of a Data Glove device and in an implementation of a small wizard game. To achieve the goal several programs and algorithms have been developed.

The implementation features developed for the thesis project are:

• Multi-thread server application for data acquisition and LAN propagation
• Client classes to read glove acquired data from the LAN
• Gestures recogniction algorithms
• Inverse Kinematic algorithms for arm animation
• Forward Kinematic algorithms for hand animation
• Simple AI algorithm for the opponent wizard
• Modeling, texturing and rigging for the game assets

Since the game relies on Zenith engine the whole project has been developed in C++. The multi-thread server application reads the data from two different devices, the Polhemus tracker and the 5DT Data Glove 14 Ultra using one thread per device.

Read data is stored and then retrived from the network-thread which generate packets for LAN propagation (using TCP/IP protocol). Several clients may connect to the server.

The game client uses the data provided from the server to animate the hand. If the hand is in pointing stance a spell starts to be casted. During the casting sequence the gesture recognition algorithm searches for any pattern to be matched. To complete a spell three patterns must be matched, and in every stage the player may choose which pattern he prefers to reproduce.
When the spell is ready it may be casted opening the hand toward the opponent.

Video

This video has been taken before Arkam being finished.
The giant monitor is from the virtual theatre I've worked on sited in the "Laboratorio di Eidomatica", University Of Milan.
The opponent wizard is still and doesn't react since the AI system was't ready yet. Gesture recognition system was almost done while some problems where still present in the data glove calibration system.
The video is anyway a good rappresentation of the work itself.