ABSTARCT :
In many application areas mobile robots need the ability to interact with people. In order to be able to communicate with even untrained users, the interaction should be as natural as possible. One of the preconditions of a natural interaction is that the robot focuses the person it is interacting with. In this paper we present a technique that combines data acquired with a laser range finder and the output of a standard face detection system to determine the positions of persons in the vicinity of the robot. We present experimental results obtained with a real robot which illustrate that our approach can efficiently detect faces of persons and that the overall processing time is reduced compared to a purely vision-based approach.
EXISTING SYSTEM :
? Existing approaches use color tracking, a mixture of features, or stereo vision to identify single persons and eventually their gestures or actions.
? Facial recognition is used by many security systems and agencies to verify the identity of a person. Many algorithms exist for face recognition, each having its own strengths and weaknesses.
? Many techniques exist to try and find the face or faces within an image. Some techniques are faster, others are more accurate.
? The last check is to try and detect whether a pair of eyes do exist in the face image or not.
DISADVANTAGE :
? The problem of detecting and tracking objects or persons has been studied intensively in the area of computer vision.
? These systems apply different algorithms to extract features and to represent the uncertainty of the robot about the position of the person in its vicinity.
? Both approaches, the purely vision-based and the purely range-data based, have their advantages and disadvantages.
? The problem with such robots is the limited payload capability resulting in reducing the usefulness of the acquired information while requiring high operation costs in terms of energy spent on flying and controlling the platform.
PROPOSED SYSTEM :
• In this paper we propose a method allowing a mobile robot to continuously focus the nearest person within a group of surrounding people with a video camera embedded in the robot’s human-like face.
• In this paper we therefore propose a combination of both sensor modalities to robustly detect persons in the vicinity of the robot.
• The robot is proposed to be used by military and security personnel in their operations.
• In a robot system is proposed in which RFID tags are used to identify authorized personnel which are allowed to control the robot.
ADVANTAGE :
? The range scanners additionally provide accurate bearing information so that sub-images containing the faces can efficiently computed.
? The segments of the images possibly containing the faces of the persons in the vicinity of the robot can be processed more efficiently than the entire image.
? In this paper we presented an approach to combine laser and vision data for robust and efficient face detection with mobile robots.
? The technique presented in this paper provides an efficient solution to the task of identifying faces in images.
? We expect that this will make person identification more robust and thus will lead to improvements of human-robot interaction.
|