3. OBJECTIVES

Gestrures represent a natural and intuitive way of interaction and communication: we gesture in order to express ideas, concepts; we gesture in order to interact with objects. The project considers the theme of human gesture recognition for interaction with information systems (virtual reality and augmented reality applications) and robotics systems. The accent is on video processing using techniques specific to image processing, pattern recognition and computer vision.

The main objective is to obtain a mathematical model, algorithms, programs and hardware arhitectures in order to achieve a natural interaction mechanism including flexible, efficient and ergonomic gestures.

We can identify the following measurable objectives:

§ design of a gesture dictionary composed of both hand postures as well as dynamic gestures (hand postures with motion trajectories; head movements), taking into account ergonomic aspects
§ real time acquisition and interpretation of hand and head gestures (using the stereoscopic view, 2 video cameras)
§ development of an embedded system for preprocessing the video information with a System on Chip architecture
§ implementation of the proposed interaction system inside a 3D virtual reality application currently under development at Laboratoire d'Informatique Fondamentale de Lille
§ implementation of the proposed system inside an augmented reality application for a smart desk
§ experimenting the interaction technology with robotics systems (the static arm type tobot Hercules developed at the University of Suceava and a mobile prototype robot developed in collaboration with Polytech'Lille, France)

The design of the gesture dictionary will contain a serie of objectives, as follows::

§ proposal for a gesture dictionary including hand postures, gesture trajectories and head movements
§ gathering of voluntary users for collecting video data containg gesture commands
§ ergonomic studies with regards to the users' capacity of interaction by the means of gesture with an information system and considerations on the aspect of complexity, naturalness, flexibility and efficiency of the proposed gestures.

Real time acquisition of hand and head movements includes the following objectives:

§ development of a real time hand detection module in video sequences with an accuracy of 90% by analyzing and correlating the visual information from 2 video cameras (the research direction is given by skin color detection methods)
§ development of a real time head detection module in video sequences with an accuracy of 90% by analyzing and correlating the visual information from 2 video cameras (methods based on the Haar features introduced by Viola & Jones, 2001)

In order to measure the performances of the hand and head detectors, the following objectives are considered:

§ build of a video data base of gesture commands collected on a training process including voluntary users
§ separate the video data base into training and testing data
§ performance measurement using the testing data base and rapporting to existing data from the literature (computing of False Positives, False Negatives, True Positives, True Negatives and the Receiver Operating Characteristic curve [Viola & Jones, 2001])

Considering the CEEX program's objectives, the project considers:

§ development of the Romanian RD system in order to accumulate new knowledge, results and high edge experience in a leading scientifical and technological domain (natural interaction with information and robotics systems) as well as assuring  the transfer of the results towards the inner social environment for the development of its competitivity
§ concentration and optimal valorification of the scientifical and technological potential existing in Romania in high edge HCI
§ supporting formation, development, integration and consolidation of the HCI domain with respect to the interanational standards and research
§ acceleration of the alignment and integration process to the reglementations of the European Union