الفهرس 
Optical flowOnly 14 pages are availabe for public view
 |  | from | 16 |  |  |
| | | from | 16 | | |
Abstract
Object tracking is an important task within the field of computer vision. Tracking is
the problem of generating an inference about the motion of an object given a sequence of
images. The proliferation of high powered computers, the availability of high quality and
inexpensive video cameras, and the increasing need for automated video analysis has
generated a great deal of interest in object tracking algorithms.
Video tracking is often divided into target representation, target detection and target
tracking. First you create a model of the target. Then you implement techniques for
detecting the object using the target model. Lastly, you use different tracking techniques
to predict the new target location given the current location.
from the study of tracking algorithms, we choose the UKF which yields to very good
results when dealing with partial and full occlusion, since the UKF predicts the position
of a moving object based on its past values. UKF is a recursive adaptive filter that
estimates the state of a dynamic system from a series of noisy measurements.
In this thesis, a comparative study has been evaluated to verify the proposed UKF with
CBWH mean shift algorithm is superior to the traditional tracking methods. The CBWH
scheme is used to represent a target. The CBWH scheme can effectively reduce
background’s interference in target localization. So CBWH can guarantee accurate
localization of the target. Then UKF algorithm has the ability to estimate the coming state
with high level of accuracy. So the proposed algorithm can effectively overcome the
problems of object occlusion as well as background variation.
The proposed face tracking algorithm is designed based on the UKF and Viola-Jones
algorithm. Viola-Jones is used to detect face and then tracking the faces using UKF
algorithm. Viola/Jones detector is most effective only on frontal images of faces and it
can hardly cope with face rotation both around the vertical and horizontal axis. The
tracking algorithm based on Viola-Jones and UKF can track even tilted, rotated faces and
further away from the camera, which is not possible in case of viola jones only. Thus the proposed face tracking algorithm is used to improve the solution of face tracking
problems.
We apply the first proposed algorithm (UKF with CBWH-MS) on face tracking and
compared with UKF with viola/jones and Camshift algorithm. The investigational results
show that the UKF with CBWH-MS is superior to UKF with Viola/Jones when the faces
tilt.