Seminar on Image based Facial Recognition System

ARUNA V 09012D4524 M.tech(SSP) JNTUCEH

INTRODUCTION
A facial recognition system is a computer application for automatically identifying or verifying a person from a digital image or a video frame from a video source.

Functional Modules:
Face image detector : Finds the locations of human faces.

Face recognizer
Feature extractor

: Determines who this person is.

: That transforms the pixels of the facial

image into a useful vector representation.

Pattern recognizer : That searches the database to find the

best match to the incoming face image.

Variations in Facial Images

A human face is not only a 3-D object, it is also a non-rigid body.

Variations in Facial Images

Camera distortion and noise

Complex background
Illumination

Translation, rotation, scaling,

Facial expression

Makeup and hair style

Face recognition scenarios

Face verification (or authentication) Face identification (or recognition) Watch list

Face verification
Face verification is a one-to-one match that compares a query face image against a template face image whose identity is being claimed.

Face identification
It is a one-to-many matching process that compares a query face image against all the template images in a face database to determine the identity of the query face.

The Watch list

Appearance variations of the same subject under different lighting conditions and different facial expressions

Inter-subject variations versus intra-subject variations. (a) and (b) are images from different subjects, but their appearance variations represented in the input space can be smaller than images from the same subject, b, c and d.

Face recognition algorithms

1. Appearance-based (View-based) face recognition.

2.

Model based face recognition.

Appearance-based (View-based) face recognition: An image is considered as a highdimensional vector, i.e., a point in a high-dimensional vector space.

Appearance-based (View-based) face recognition

Vector representation of images Linear (subspace) Analysis PCA ICA LDA Small Sample Size Non-linear (manifold) Analysis

Kernel PCA

Appearance-based (View-based) face recognition: Linear (subspace) Analysis

PCA: The Eigen face algorithm uses the Principal Component Analysis ICA:

Eigenvectors corresponding to the 7 largest Eigen values, shown as p x p images, where p x p = n

Appearance-based (View-based) face recognition:

Linear (subspace) Analysis
LDA: Both PCA and ICA construct the face space without using the face class (category) information. The whole face training data is taken as a whole.

In LDA the goal is to find an efficient or interesting way to represent the face vector space.

Feature-based Elastic Bunch Graph Matching: Bunch Graph:

Multi view faces overlaid with labelled graphs

Elastic Graph Matching:

To identify a new face, the face graph is positioned on the face image using elastic bunch graph matching

Feature-based Elastic Bunch Graph Matching

AAM - A 2D Morphable Model

AAM Fitting

Feature-based Elastic Bunch Graph Matching

3D Morphable Model Model Fitting

Recognition

Top row: initial parameters, Middle row: Results of fitting, rendered on top of the input images. Bottom row: Input images. The fifth column is an example of a poor fit.

Applications
Identity verification for physical access control in buildings or security areas

To allow secure transactions through the Internet, face

verification may be used instead of electronic means like passwords or PIN numbers Face identification has also been used in forensic applications for criminal identification

Thank you