VIVEKANANDHA COLLEGE OF ENGINEERING FOR WOMEN (Department of ECE)HEAD GESTURE RECOGNITION FOR HANDS – FREE CONTROLOF AN INTELLIGENT WHEEL CHAIR Presented by : Suganya D (III ECE) Suganthi priya T(III ECE)
OBJECTIVE This paper presents a novel hands-free controlsystem for intelligent wheelchairs (IWs) based onvisual recognition of head gestures for elderly anddisabled people who have restricted limbmovements
ABSTRAC T Electric-powered wheelchairs (EPWs) have been rapidly deployed over the last 20 years These EPWs are controlled by users’ hands and are very difficult for elderly and disabled users. As cheap computers and sensors are embedded into EPWs, then they named as intelligent wheelchair(IWs).
INTRODUCTION Our IWS is based on novel head gesture-based interface (HGI), namely RoboChair, Based on the integration of the Adaboost face detection algorithm and the Camshift object tracking algorithm. Head gesture recognition is conducted by means of real-time face detection and tracking.
SYSTEM HARDWARE STRUCTURE : Consists of parts as follows. six ultrasonic sensors at a height of 50 cm. DSP TMS320LF2407-based controller. a Logitech 4000 Pro Webcam. a local joystick controller. Intel Pentium-M 1.6G Centrino laptop
CONTROl SYSTEM Of ROBO CHAIR Control system able to achieve both real time signal processing and high performance driving control due to the following features viz., Excellent processing capabilities(30 MIPS) Compact peripheral integration Two control modes of robo chair: Intelligent control mode Manual control mode
MANUAl CONTROl MODE In this mode of operation, Robochair is controlled by the JOYSTICK JOYSTICK is connected to an A/D converter of the DSP motion controller.
INTEllIgENT CONTROl MODE – Robochair is controlled by the proposed ( Head Gesture Interface ) HGI. – A Logitech web camera is used to acquire the facial images of the user. – Image data is sent to the laptop. Head gesture analysis and decision making stages are implemented. – Finally, the laptop sends control decision to the DSP motion controlled that actuates two DC motors.
HgI ( HEAD gESTURE INTERfACE ) It uses two algorithm. Adaboost face detection algorithm Camshift object tracking algorithm ADABOOST FACE DETECTION ALGORITHM ADVANTAGES: Extracts the Haar-like features of images that contain image frequency information. Adaboost is able to detect profile faces High accuracy and speed in face detection CAMSHIFT OBJECT TRACKING ALGORITHM ADVANTAGES: Very efficient color tracking method based on image hue and achieve real time performance.
INTEgRATION Of BOTH AlgORITHMS Since low cost IW’s have limited onboard computing power, Adaboost face detection algorithm can’t achieve real time performance. On the other hand, camshift face tracking algorithm runs very fast ,but is not robust to varying illumination conditions and noisy backgrounds. So to obtain both speed and accuracy, it is necessary to integrate both algorithm.
HEAD gESTURE RECOgNITION To recognize the head gesture ,Adaboost frontal, left profile and right profile head gesture classifiers are adopted. If the profile face is detected, our Robochair is going to turn left or right. By calculating the precise nose position can detect the exact frontal face head gesture using classical template matching method.
NOSE TEMPlATE MATCHINg There are five frontal head gestures to be recognized, namely: 1. center frontal; 2. up frontal; 3. down frontal; 4. left frontal; and 5. right frontal.
ROBOCHAIR ACTIONS fOR MOTION CONTROl COMMANDSRules to be followed for action for Robochair: Speed up(if frontal face up is recognized) Slow down until stop(if frontal face down is recognized) Turn left (if left profile/frontal face is recognized) Turn right (if right profile/frontal face is recognized ) Keep speed (if central face is recognized)
DEMONSTRATION fOR PROfIlE fACES A sequence of images under head gesture control are Turn right Right up Turn left Turn left with hand color noise
CONClUSION This paper describes the design and implementation of a novel hands-free control system for IW’s. A robust HGI, is designed for vision-based head gesture recognition of the Robo Chair user. To avoid unnecessary movements caused by the user looking around randomly, our HGI is focused on the central position of the wheelchair
REfERENCES: Bradski, G. (1998), “Real-time face and object tracking as a component of a perceptual user interface”. Ding, D. and Cooper, R.A. (1995), “Electric powered wheelchairs”, IEEE Control Systems, Galindo, C., Gonzalez, J. and Fernandez- Madrigal, J.A. (2005), “An architecture for cognitive human-robot integration. Application to rehabilitation robotics”, Proceedings of IEEE International Conference on Mechatronics