Description

ABSTRACT

Face recognition is successful by using Gabor filter and neural networks. In this paper we present a new technique of face detection and recognition in color images. In this 40 different Gabor filters are applied on an image will result 40 different images with different orientations. This paper addresses a novel algorithm in order to detect face features and extract their corresponding geometric points. In those 40 filtered images maximum intensity points are calculated and mark them as fiducially points. To reduce those fiducially points distance between those points is considered. Here we are using feed forward neural network to classify the query image with database images. Finally we will get the results like the face is matched or not by estimating the features of input images. Particularly we are using PCA(principal component analysis) to extract the features.

DEMO  VIDEO

INTRODUCTION

Now a day, many applications used by the civilians and army or police forces require effective face recognition. In this case face recognition is very useful to easily detect the human faces. This face recognition is a very challenging area in computer vision and pattern recognition due to various variations in facial expressions, poses, Illumination, changes in temperature, sweating, aging and skin discoloration. If there are any changes due to injury, fashion etc there is further complexity. Most of the methods are based on neural networks, feature extraction, skin color, and others are based on template matching. Many techniques for face recognition are developed. There are two approaches: geometrical and general methods. Geometrical methods are based on measures extracted between the facial features such as eyes, mouth, nose and chin. In general methods total face is treated as a single. For face detection, we can use PCA, Eigen faces, neural filters and Gabor filters .The points marked on the face by Gabor and neural networks can be compared by pre stored data base of faces. If match occurs, then the face will be detected. If we want high level of security then we can use face detection, cards checking etc. This type of face recognition system is used in areas such as air ports for the official people and the staff for direct sending them rather than normal people. For any transactions over the internet, we have to remember PIN numbers and password; it is very difficult so, face verification is implemented. It is very useful because pin numbers and passwords can be hacked by any one. Face recognition can be used for identification and surveillance on secure areas or on public places to detect illegal activities. It can also be used in government security issues like national ID cards, passports and license, etc. Neural networks are used for face recognition and used for decision making to perform complex functions. In FFNN neurons form the layers. In the training process, the weights that connect neurons of consecutive layers are obtained using the inputs of neural network and the desired output.

EXISITNG SYSTEM

Normally while estimating the features of an image we are using general feature extraction algorithm. In existing system we approaching the results by using SVM(support vector machine) classifier and to calculate the features of image we are using GLCM features. The morphological filters were used in this project but we have some draw backs in existing system.

PROPOSED METHOD

In the abstract the algorithms which we are using proposed method has been given. We are proposing Gabor filters and FFNN(Feed Forward Neural Network) to classify the images. PCA algorithms we are using in different types of applications

BLOCK DIAGRAM

ADVANTAGES

• Advanced feature extraction
• High accuracy of calculation

APPLICATIONS

• Real time object recognition
• Image compression

SOFTWARE REQUIREMENTS

• MATLAB 2014 or above versions

MATLAB GUI FOR FACE RECOGNITION USING GABOR FEATURES

CONCLUSION

This paper proposes a classification-based face detection method using Gabor filter features. Considering the desirable characteristics of spatial locality and orientation selectivity’s of the Gabor filter, we design filters for extracting facial features from the local image. The feature vector based on Gabor filters is used as the input of the classifier, which is a Feed Forward neural network (FFNN) on a reduced feature subspace learned by an approach simpler than principal component analysis (PCA).

REFERENCES

1]. S. Khan fir, Y. B. Jemaa, “Automatic facial features extraction for face recognition by neural networks”, 3rd International Symposium on Image/Video Communications over fixed and mobile networks (ISIVC), Tunisia 2006.

[2]. R. L. Hsu, M. A. Mottaleb and A. K. Jain, “Face detection in color images”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 24,No. 5, pp. 696-706, May 2002.

[3]. W. Zhao, R. Chellappa, P. J. Phillips, and A. Rosenfeld.“Face recognition: A literature survey. ACM Computing Surveys”, 35(4): pages 399–458, 2003.

[4]. A.J.Goldstein, L.D. Harmon, and A.B Lesk, “Identification of Human Faces,” Proc. IEEE, May 1971, Vol. 59, No. 5, 748-760.

[5]. H. A. Rowley, S. Baluja, T. Kanade, “NeuralNetwork-Based Face Detection”, IEEE Trans. On Pattern Analysis and Machine Intelligence, vol.20, No. 1, Page(s). 39-51, 1998.