face detection for access control

12
FACE DETECTION FOR ACCESS CONTROL By Dmitri De Klerk Supervisor: James Connan

Upload: earl

Post on 25-Feb-2016

32 views

Category:

Documents


2 download

DESCRIPTION

Face Detection for Access Control. By Dmitri De Klerk Supervisor: James Connan. Introduction. Project description Implementation of a face detection system that may be used for access control. Face detection (done prior to face recognition) - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Face Detection  for  Access Control

FACE DETECTION FOR ACCESS CONTROL

By Dmitri De Klerk

Supervisor: James Connan

Page 2: Face Detection  for  Access Control

INTRODUCTION Project description

Implementation of a face detection system that may be used for access control.

Face detection (done prior to face recognition) Finds the locations and sizes of human faces in

arbitrary (digital) images.

Motivation for face detectionNo need to position your face inside a fixed

size box.

Page 3: Face Detection  for  Access Control

USER REQUIREMENTS Expected from a solution:

Detect a persons face within an image. frontal face poseminor variations in lighting conditionsminor variations in facial expressions

Not expected from a solution:non frontal face poseextreme lighting conditions: darkness

or too much lighttotal or partial occlusion of face

Page 4: Face Detection  for  Access Control

REQUIREMENT ANALYSIS Proposed Solution

Detect faces in image. Pass the biggest detection to the recognition

face. User closest to the camera.

Resize the users face to the standard recognizable size.

Recognize the users face Check whether the user is authorized.

Grant or deny the user access.

Page 5: Face Detection  for  Access Control

ANALYSIS & INTERFACE

Page 6: Face Detection  for  Access Control

DESIGN & IMPLEMENTATION Face Detection

Viola and Jones approach Adaboost Haar Features Cascade classifiers

Face Recognition Neural network [Desmond Van Wyk 2006]

Page 7: Face Detection  for  Access Control

DETECTOR Training Data

3398 hand labeled faces All frontal Took 7 days to train

607 non faces different scale Faces are normalized

Scale, translation Many variations

Across individuals Illumination Pose (rotation both in plane and out)

Page 8: Face Detection  for  Access Control

PERFORMANCE TESTING MIT+CMU frontal face test set.

Images collected at CMU and MIT.

275 Correctly Detected out of 472 images. 58% Detection rate

Page 9: Face Detection  for  Access Control

TOOLS AND LANGUAGES Java Media Framework (JMF)

Capturing images and communicating with camera

NetBeans IDE Java Programming Languages

Page 10: Face Detection  for  Access Control

FACE DETECTION DEMO

Page 11: Face Detection  for  Access Control

REFERENCES AdaBoost

http://en.wikipedia.org/wiki/AdaBoost

Builds on the work of Desmond Van Wyk Face Recognition System for Access Control (2006)

How face detection work http://www.cognotics.com/opencv/servo_2007_series/part_2/sid

ebar.html

Fast Face Detection Using AdaBoost Julien Meynet [16 July 2003]

The Boosting Approach to Machine Learning – An overview Robert E. Schapire [19 December 2001]

Page 12: Face Detection  for  Access Control

QUESTIONS AND ANSWERS Thank you!