For the second milestone the system will be completed with a first prototype of three modules: the acquisition module, the range data compression module and the coloring model.
Acquisition. The acquisition of range data will be carried out by projecting on the object's surface a moving grid and surveying the grid images through a stereo-camera. The system is already under development for face reconstruction and some images are reported here.
Data compression. The amount of data points, which can be collected in the acquisition phase, can be in excess for any visualization or processing system. We will explore algorithms for intelligent decimation of the data, which can also clean the acquisition noise. These algorithms have to work in a few time and produce sets of small cardinality. A software prototype in C with OpenGL interface will be produced.
Coloring. We suppose that not all the components of our system will be endowed with texture mapping modules. Therefore we will develop procedure for automatically or semi-automatically produce the color field for the models reconstructed through HRBF networks (milestone I), in the form of Gouraud shading. A software prototype in C with OpenGL interface will be produced.