8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 1/10

 

Neural Based:

Real time face detection &Recognition using haar

features

Second Presentation

Submitted By:Amit Shah (07-CSS-62)

Sheel Sindhu Manohar(07-CSS-62)

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 2/10

 

Introduction

Describes a face detection framework that is capable of processingimages extremely rapidly while achieving high detection rates.

Feature Based Detection

Image representation using “Integral Image” which allows the

features used by our detector to be computed very quickly.

Built using the AdaBoost learning algo-rithm (Freund and

Schapire, 1995) to select a small number of critical visual features

from a very large set of potential features.

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 3/10

 

Feature Used

The simple features used are reminiscent of Haarbasis functions

which have been used by Papageorgiouet al. (1998).

We used three kinds of features.

The value of a two-rectangle feature is the difference between thesum of the pixels within two rectangular regions. The regions have

the same size and shape and are horizontally or vertically adjacent

  A three-rectangle feature computes the sum within two outside

rectangles subtracted from the sum in a center rectangle.

Finally a four-rectangle fea- ture computes the difference between

diagonal pairs of rectangles

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 4/10

 

Haar Feature

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 5/10

 

Integral Image

Rectangle features can be computed very rapidly using an

intermediate representation for the image which we call the

integral image. The integral image at location x, y contains the

sum of the pixels above and to the left of x, y, inclusive:

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 6/10

 

Features detection

We choose to generate a very large and varied set of rectangle

features. Typically the representation is about 400 times

overcomplete.

Rectangle features provide a rich image representation which

supports effective learning.

The extreme computational efficiency of rect-angle features provides

ample compensation for their limitations.

Like most face detection systems, our de-tector scans the input at

many scales; starting at the base scale in which faces are detected

at a size of 24 × 24 pixels, a 384 by 288 pixel image is scanned at

12 scales each a factor of 1.25 larger than the last.

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 7/10

 

Feature detection(contd.)

The conventional approach is to compute a pyramid of 12 images,

each 1.25 times smaller than the previous image. A fixed scale

detector is then scanned across each of these images.

Computation of the pyramid, while straightforward, requires

significant time. Imple-mented efficiently on conventionalhardware (using bi-linear interpolation to scale each level of the

pyramid) it takes time according to the processor(System

dependent).

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 8/10

 

Face detection in brief 

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 9/10

 

Wor !ill No" #one

Face has been detected succesfully.

AdaBoost is an effective pro-cedure for searching out a small number

of good “fea-tures” .

We are working on the Training part of the algorithm. Database Used For Training: http://cbcl.mit.edu/software-

datasets/heisele/download/MIT-CBCL-facerec-database.zip

8/13/2019 Neural Based Face Detection

http://slidepdf.com/reader/full/neural-based-face-detection 10/10

 

Future $ddition

Extracting faces detcted in face detection. And storing in a database

which would be used for further training of the classifier.

Real time Face recognition