5. PROJECT JUSTIFICATION:

5.1. Project relevancy for the thematic sector and the way the project matches the national research policy

By approaching the very modern and up to date domain of gesture based interaction with information and robotic systems, the proposal unifies in common bounds the specific and specialization of the research previously conducted by the national partners as well as the europeen collaborators for the domains of video processing, pattern recognition, virtual and augmented reality applications and robotic systems. The project with its results will contribute semnificatively to reaching the compatibibility and competitivity levele required for the integration in the European research area as well as for the preparation for participating at the CDT 7 research of the European Union.

The proposal is part of the CEEX domain Information and communication technologies, with the thematic area „Robotic systems: advanced autonomous szstems; cognition, control, action skills, natural interactions; miniaturisation" by considering natural gesture based interaction for the interfacing with information systems (such as virtual and augmented reality applications) and for robotic systems. We are aiming at a natural, flexibile and efficient interaction by considering gestures as a natural and intuitive way of interaction and communication.

We also consider the proposal to be relevant under the specific requirements in the scenario of the integration in the European research area and for reaching a compatibility and competitivity level for the national research with regards to the European Union specifications. Considering the possible applications for this reaserch, the project theme matches the thematical areas promoted by two important European research programs:

§ the PC7 research program of the European Union for 2007-2013 with regards to information and communication technologies
§ the initiative of the European Union, the i2010 European Information Society 2010.

We estimate the proposal for a gesture based interaction system with information and robotics systems viable and duable with regards to the plan of development, an affirmation that is motivated by the current achievements and research directions of the partners:

§ the research being done at the University of Suceava in the period 1992-2006 with regards to the controle of robotics systems, with the specific implementation of the robotic arm Hercules
§ the PhD thesis having research themes specific to computer vision, virtual reality systems, video processing and robotics systems. To be noticed the collaboration in the domain of gesture based interaction with virtual reality systems with LIFL of Universite des Sciences et Technologies de Lille with regards to the codirected PhD thesis (PhD student Radu Daniel Vatavu, directors: Şt.Gh.Pentiuc (University of Suceava), Christophe Chaillou (Universite des Sciences et Technologies de Lille): Real time acquisition of human gestures for interaction with virtual reality, started in 2004; the results of the projects regarding robotics controlling during the Socrates collaboration between University of Suceava and Polytech'Lille)
§ research conducted at the Computer Science Research Center of the University of Suceava with regards to advanced systems for process control, artificial intelligence, distributed systems.

As risks with regards to the implementation of the gesture based interaction system, we have to consider the riscs that are specific to the computer vision applications in general. There is always a limit as to the performances of such applications. We are targeting a performance percent of 90% in what concerns the accuracy of hands detection and head's movements in video sequences.

Risk	Alternative solution
Detection with insufficient accuracy of hand postures may be caused by video camera's resolution and speed of gesticulation	We shall reconsider the processing at a higher resoltution with a dedicated video camera that will follow only hand movements. Considerations will be made on the optimal positioning of this new camera and with regards to the advices for speed gesticulation addressed to the user.
The time necessary for real time processing considering the amount of visual information from 2 video cameras	We shall consider a pseudo real time interaction by analyzing 15 frames per second instead of 30 (15 fps are sufficient for inducing the feeling of real time interaction to the user).