VISITOR IDENTIFICATION APPLYING DIGITAL IMAGE PROCESSING INTRODUCTION INTRODUCTION

VISITOR IDENTIFICATION APPLYING DIGITAL IMAGE PROCESSING

INTRODUCTION

INTRODUCTION In the year of 2015, a man was accused of masquerading as a doctor Unfamiliar Face Wrong Decision by the human-being

OBJECTIVE

OBJECTIVE 1. 2. 3. 4. 5. 6. Application of Face detection Interface of computer and surveillance system Create the database for the faces Face Identification system User Interface RFID Card System

BACKGROUND

BACKGROUND Face Detection � Recognize face by computer Surveillance System (Webcam) � Capture the video � Cheap Microsoft Visual Studio � Visual Basic � User Friendly � Easy to Use

BACKGROUND Open. CV �A Library for the face Detection � Convenient to use � C++ Syntax Emgu. CV � Platform to use Open. CV in Visual Basic Arduino (MEGA 2560) � Easy to program � User Friendly � Many Library provided MFRC 522 RFID Reader/Writer Card

WORKING PROCEDURE

STUDYING THE FACE DETECTION Many Method Triangle Portion Method Visual Basic Open. CV Using Emgu. CV

CREATING FACE DETECTION SYSTEM Fundamental Face Detection Created � XML provided by Open. CV to execute face detection � Face found in the picture will be surrounded by square Start face Detection in Video Real Time. Video Image Face Detection System

CREATING FACE DETECTION USING WEBCAM Revolution is high now Visual the human face obviously.

CREATING FACE DETECTION USING WEBCAM Internal Function of Open. CV � Convert the video to frame - img. Original = cap. Webcam. Query. Frame()

CREATING FACE DETECTION USING WEBCAM Internal Library of Microsoft. � Function to read serial port message Imports System. IO. Ports Face Detection is finally successfully constructed

CREATING THE DATABASE SYSTEM FOR FACE IDENTIFICATION AFTERWARD Button Added RF ID must be inputed Record can be saved with face detected

CREATING THE DATABASE SYSTEM FOR FACE IDENTIFICATION AFTERWARD Button Added RF ID must be inputed Record can be saved with face detected

HIDING IDENTIFICATION NUMBER AND RECORD LOCATION Privacy, Security Hiding RF ID

FACE IDENTIFICATION USING “NORM” FUNCTION OF OPENCV Internal Function of Open. CV � Norm() � Compare difference of two pictures The repeatability is low Result Affected � User hair style � Light level

FACE IDENTIFICATION USING RATIO AND ANGLE BETWEEN EYE, NOSE AND MOUTH Different Face � Different angle between eyes, nose and mouth � Ratio of distance of eyes to nose and distance of eye to mouth

EYE, NOSE AND MOUTH DETECTION BEFORE FACE MATCHING Saved Record without eyes, nose and mouth detected Checking mechanism will be then failure Ensure checking mechanism is successfully

REDUCE THE FACE REGION TO CONDUCT THE FACE SIMILARITY CHECKING Improve the checking mechanism Cut out the face � Remove the hair Result become better

RF ID SYSTEM CONSTRUCTION Face Detection already consume many resource Wait the serial port to response Defer the speed of whole program It takes long time on processing…. Program Start Read the RF ID Face Detection

MULTITHREADING OF THE RF ID SYSTEM Define RF ID System as thread RF ID System runs separately � It would not affect the main program

WORKING PROCEDURE Eye, Nose and Mouth Detection before face matching Reduce the Face Region to conduct the face similarity checking RF ID System Construction Multithreading of the RF ID system

APPLICATION

RECORD LOCATION

SAVE THE RECORD

DISPLAYED RF ID

RECALL THE SAVE RECORD

FACE SIMILARITY CHECKING

FURTHER DEVELOPMENT The Cost of RF ID card Reader � Replaced with alternative Card Reader Accuracy of Face Detection and similarity checking � More Face feature can be used for calculation � Infra-Red May also help Replace Traditional Personal Computer with Raspberry � Raspberry is also microprocessor � The cost of PC is high � Raspberry relatively checper

Q&A

END