AdaBoost人脸检测算法DSP实现.pdf

成都理工大学硕士学位论文

Implementation of face detection algorithm for AdaBoost
base on DSP
Introduction of the author:Yuquanhe, male, was born in August, 1984 whose
tutor was associate Professor Zhu graduated from Chengdu University
of Technology munication and information system major and was granted the
Master Degree in June, 2010.
Abstract
Face detection is to determine whether there existsin a picture or continuous
video ,if have, marked the size and location of the is one of the most popular
comput-er pattern recognition research topics,is applied in video conference, remote
monitoring, remote treatment. Many scholars,and experts have done a lot a lot of
work and achieved outstanding results and give a lot of effective
mostly achiev- ed in the PC,and can not meet the requirement of convenient, real-time
and low cost .
Under this situation,this paper put forward to achieve Adaboost face detection
algorithm in the DSP environment, The main work of this paper is as follow:
1. Face detection algorithms are mostly, after understanding the relevant
algorithms pareing their ,I chose Adaboost face detection algorithm and Its
features is adaptable, small error rate. If the training samples is too good
enough,cascade classifier based haar can make the error rate reduce to arbitrarily
small, can be adapted to plex environments.
2. The paper use MATLAB to achieve Adaboost face detection algorithm,we can
focus on the performance of algorithm,but not study how to achieve on their own.
the 6000 DSP chips provide strong hardware suppor in the video image
developed. DSP / BIOS real-time system programming divided into three tasks: as
image processing, image analysis, image display,task munication between
each other.
is open puter vision library, a powerful tool for image
processing, image engineers can significantly reduce product development time,The
paper researc