Your SlideShare is downloading. ×
face recognition system
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×

Introducing the official SlideShare app

Stunning, full-screen experience for iPhone and Android

Text the download link to your phone

Standard text messaging rates apply

face recognition system

1,376
views

Published on


0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
1,376
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
83
Comments
0
Likes
1
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. Face recognition: Face recognition is one of the most importantbiometric which seems to be a good compromise between actualityand social reception and balances security and privacy well. Facerecognition system fall into two categories: verification andidentification. Face verification is a 1:1 match that compares a faceimages against a template face images, whose identity is beingclaimed. On the contrary, face identification is a 1: N problem thatcompares a query face image against all image templates in a facedatabase. Face localization, feature extraction, and modeling are themajor issues in automatic facial recognition.
  • 2. Facial recognition is an effective biometricattribute/indicator. Different biometric indicators are suited fordifferent kinds of identification applications due to their variations inintrusiveness, accuracy, cost, and ease of sensing. Among the sixbiometric indicators considered facial features scored the highestcompatibility.
  • 3. Comparison of various biometric features: (a) based on zephyr analysis (b) based on MRTD(Machine Readable Travel Documents)compatibility
  • 4. :In the 1970s,Goldstein,Harmon,and Lesk1 used 21 specificsubjective markers such as hair color and lip thickness to automatethe recognition. The problem with both of these early solutions wasthat the measurements and locations were manually computed. In 1988,Kirby and Sirovich applied principle componentanalysis, a standard linear algebra technique, to the face recognitionproblem. This was considered somewhat of a milestone as it showedthat less than one hundred values were required to accurately code asuitably aligned and normalized face image.In1991, Turk and Pent land discovered that while using theeigenfaces techniques, the residual error could be used to detectfaces in images a discovery that enabled reliable real-time automatedface recognition systems.
  • 5. A framework for face recognition based attendancesystem:
  • 6. Predominant Approaches:There are two predominant approaches to the face recognitionproblem: 1.Geometric (feature based) 2.Photometric (view based)
  • 7. As researcher interest in face recognition continued, Many differentalgorithms were developed, three of which have been well studied inface recognition literature: 1. Principal Components Analysis (PCA) 2.Linear Discriminant Analysis (LDA) 3.Elastic Bunch Graph Matching (EBGM)
  • 8. Principal Components Analysis (PCA) :•PCA, commonly referred to as the use of eigenfaces, this techniquepioneered by Kirby and Sirivich in 1988.•In PCA, the probe and gallery images must be the same size and mustfirst be normalized.•The PCA approach is then used to reduce the dimension of the data bymeans of data compression basics and reveals the most effective lowdimensional structure of facial patterns. This reduction in dimensionsremoves information that is not useful and precisely decomposes the facestructure into orthogonal (uncorrelated) components known aseigenfaces.
  • 9. • Each face image may be represented as a weighted sum (feature vector) of the eigenfaces, which are stored in a 1D array.• A probe image is compared against a gallery image by measuring the distance between their respective feature vectors.• The PCA approach typically requires the full frontal face to bepresented each time otherwise the image results in poorperformance.•The primary advantage of this technique is that it can reduce the data needed to identify the individual to 1/1000th of the data presented
  • 10. Linear Discriminant Analysis (LDA): LDA is a statistical approach for classifyingsamples of unknown classes based on training samples with knownclasses. This technique aims to maximize between-class (i.e., acrossusers) variance and minimize within-class (i.e., within user) variance. Inthis each block represents a class, there are large variances betweenclasses, but little variance within classes. When dealing with highdimensional face data, this technique faces the small sample sizeproblem that arises where there are a small number of available trainingsamples compared to the dimensionality of the sample space.
  • 11. Elastic Bunch Graph Matching (EBGM): EBGM relies on the concept that real face images have manynon-linear characteristics that are not addressed by the linear analysis methods . suchas variations in illumination(outdoor lighting vs. indoor fluorescents), pose (standingstraight vs. leaning over) and expression (smile vs. frown). A Gabor wavelet transformcreates a dynamic link architecture that projects the face onto an elastic grid. The Gaborjet is a node on the elastic grid, notated by circles on the image below, which describesthe image behavior around a given pixel. It is the result of a convolution of the imagewith a Gabor filter, which is used to detect shapes and to extract features using imageprocessing. Recognition is based on the similarity of the Gabor filter response at eachGabor node. This biologically-based method using Gabor filters is a process executed inthe visual cortex of higher mammals. The difficulty with this method is the requirementof accurate landmark localization.
  • 12. Elastic Bunch Graph Matching
  • 13. :1. Digital Image Processing Using MATLAB by Rafael C.Gonzalez, Richard E.Woods , Steven L.Eddins.