Upload
ramesh-akula
View
32
Download
0
Embed Size (px)
Citation preview
Put our project title as it is
Project Guide Mr.JanaBhaskaraRaoM.E
Asst.Professor
Submitted byM.Srilatha 680552068C.Sai Ram 680552020J.Venkatesh 680552040 K.Ram Nikesh 680552051K.Jagadeesh Babu 680552057
Dept of Electronics and Communication EngineeringAnil Neerulonda Institute of Technology and Sciences
ABSTRACT Face recognition is a biometric identification technology, which
has the most potential. Research on face recognition technology has a great theoretical
and applied value. A new face recognition algorithm for LSI implementation
suitable for embedded applications is implemented. Although many algorithms are there to recognize a face recently
suggested has a relatively stable performance under the variety of environmental disturbances, the problem still lies on computational cost as well as memory usage.
In this project we will propose as algorithm based on a Pseudo fisher face matrix which is derived from generic data sets and down sampled Gabor features, which reduces these costs.
We will compare performance of the proposed method with a traditional one based on Eigen face through FERET data base.
An experimental implementation demonstrated the proposed algorithm draws significant performance rations. The project work is to be implemented using MATLAB.
INTRODUCTION
An automated face authentication system has been an active research subject for several decades.
A face authentication system consists of two face processing tasks. They are
1. Face detection
2. Face recognition Gabor feature extraction is one of the most popular methods
for face recognition. Gabor-Fisher classifier (GFC) method is a combination of
Gabor wavelets and Fisherface.
Gabor Pseudo Fisherface classifier
. . . . 1
Image1
Image2
Image n
preprocessing
Normalized image of Image1
Normalized image of image2
Normalized image of image n
Generic training dataset
Gabor Extract / Down sample
Linear Discriminant Analysis
conclusion This method will improves memory usage as well as
computational cost, it will reduces data size regarding the feature per person