Face Tracking Using Webcam

Real Time Face Tracking Using Web Camera

Abstract

This paper, presents a real time object tracking system under various conditions such as changes in lighting, object speed and background. This system is developed in order to achieve a robust, low cost surveillance system that can replace the more expensive CCTV systems available. Many of these devices provide fixed, poor, low-resolution images that are not very reliable and cannot be used for further applications. We have implemented a real time object tracking system using a web camera mounted on a pan-tilt base that is able to perform horizontal and vertical motions. Experiments have shown good tracking results for most conditions, except in poor lighting conditions and when tracking fact moving objects.

1 Introduction

Among the many fields of computer usage, one that is gaining much attention over the last two decades is remote monitoring and surveillance [1]. Computers are now being used in places where it is difficult or unsafe to place a human being.

Remote surveillance and security monitoring devices are becoming increasingly popular in the world today. Remote surveillance is widely used in the monitoring of restricted areas, hazardous environments, unfriendly territory, and many others. With the recent threats to security in countries throughout the world, it is also desirable to have remote monitoring devices in places such as airports, government buildings, prominent buildings and so on.

Real-time Object Tracking refers to the use of special purpose computer vision hardware and software to follow up on the motion of objects detected in dynamic scenes at rates which are high enough to be used in surveillance systems or decision-making, in real-world situations. Many applications have been developed for monitoring public areas such as offices, shopping malls or traffic highways. In order to control normal activities in these areas, tracking of pedestrians and vehicles play an important role in video surveillance systems.

A number of researchers have carried out a considerable amount of work in the area of object tracking. A large number of this work concentrate on tracking of objects in a video stream. There is still a lot of room for improvement in the area of real time object tracking in a real world environment.

Nelson [2] addresses the problem of identifying independently moving objects from a moving observer - an active vision system. He proposes two methods, one making use of information about the motion of the observer and the second using knowledge about how certain independently moving objects move. The first method, known as constraint ray filtering, is based on the idea that in any rigid environment, the projected motion of any point is constrained to lie on a one dimensional locus in the velocity space whose parameters depend only on the motion of the observer and the location of the image point. The second technique, called animate motion, uses the concept that the motion of the observer is generally slow and smooth, whereas the apparent motions of independently moving objects are comparatively changing more rapidly.

Betke et al [3] have developed a system that recognises and tracks multiple vehicles from sequences of gray-scale images taken from a moving car in ``hard'' real-time. Recognition is accomplished by combining the analysis of single frames with that of motion information provided by multiple consecutive frames. In single frames, cars are recognised by matching deformable gray-scale templates after detecting image features such as corners and by evaluating how those features relate to each other. They are also recognised by differencing consecutive frames and by tracking motion parameters typical for cars.

Sinclair et al [4] made use of image velocity in deriving their segmentation algorithm. They offer a novel means of segmenting independently moving objects from rigid backgrounds. One of the constraints of their system is that the camera motion can be arbitrary with the proviso that independently moving objects are not moving parallel to it. Their method relies on describing the motion of points in the world in terms of their angular velocity relative to the camera. In the case of a static scene, the method allows the recovery of the direction of the camera motion and its angular velocity from the optical flow and its first derivatives at only three points.

2. Object Tracking System Implementation

Figure 1 General Layout of the Face Tracking

Figure 1 shows the general layout of the face tracking system. It consists f two main parts: the motion tracking and the face recognition. The motion tracking part consists of both hardware and software.

The hardware for motion tracking is made up of two stepper motors and their corresponding drivers and the software is designed to detect motion and activate the stepper motor to follow the said motion.

The object tracking system consists of four main modules. They are the motion detection module, template extraction module, motion template matching module and the tracking module. The flowchart of the motion tracking system is shown in Figure 2.

Figure 2 Overview of the Tracking System

In order to efficiently detect motion within the viewing area of a camera, the motion detection module of the motion tracking system requires a threshold value. This threshold value is determined within the system initializer function and can be calculated by computing the sum of differences between two consecutive still images.

Threshold value = SS ( кIprevious[i][j] - Icurrent [i][j] к) .................(1)

Where I and J are the numbers of columns and rows of the image respectively. The threshold value obtained from the above formula is used in the motion detection module.

The motion detection module is used to detect motion produced by any moving object within the viewing area of the camera. The input is a sequence of images captured by the camera, and the output states whether or not motion has been detected. Due to computation cost-effectiveness,