HUMAN FACE DETECTION SYSTEM
ON DIGITAL IMAGE
(Thesis Proposal)
presented by
NamaMhs
NIM: XX.YY.ZZZ
To
INFORMATICS ENGINEERING DEPARTMENT
STMIK STIKOM BALIKPAPAN
APPROVAL SHEET
Thesis Proposal by title
DETECTION SYSTEM
HUMAN FACE ON DIGITAL IMAGE
presented by
NamaMhs
NIM: XX.YY.ZZZ
been approved by the Department of Information STIKOM Balikpapan with lecturers:
1. ......................................................................
2. ......................................................................
Balikpapan, dated ........................
Chairman of the Department of Information
Setyo Nugroho, ST, MKom
DETECTION SYSTEM IMPLEMENTATION
HUMAN FACE ON DIGITAL IMAGE
1. BACKGROUND
Nowadays more and more facial recognition technology is applied, among others, for biometric recognition system (which can also be combined with other biometric features such as fingerprints and voice), search and indexing system on a database of digital images and digital video databases, security systems access control area limited, video conferencing, and human interaction with computers.
In the field of face processing research (face processing), the detection of human faces (face detection) is one of the very important initial step in the process of facial recognition (face recognition). Face recognition system is used to compare an input face images with a face database and generate a face that best matches the image if any. While the face authentication (face authentication) is used to verify the authenticity / similarity of a face with a face data that has been entered previously. Areas of research that are also associated with face processing is localized face (face localization) is face detection, but with the assumption that there is only one face in the image, the face tracking (face tracking) to estimate the location of a face in a video in real time, and facial expression recognition (facial expression recognition) to recognize human emotions (Yang, 2002).
In certain cases such as photo shoots for ID card, driver's license, and credit cards, the images obtained generally contains only one face and have a uniform background and lighting conditions prearranged so that the face detection can be done more easily. But in other cases often obtained images containing more than one face, has a varied background, which is not necessarily lighting conditions and faces varying size in the image. An example is the image obtained at the airport, terminal, building entrances, and shopping centers. In addition to images obtained from photographs in the media or video recordings. In the general case of faces in the image has a shape very varied backgrounds.
This study will focus on the problem of face detection. With face detection system that is accurate, then the next process that face recognition can be done more easily.
2. PROBLEM FORMULATION
Face detection problem can be formulated as follows: the input is an arbitrary digital image, the system will detect if there are human faces in the image, and if there is then the system will tell how many faces are found and where is the location of the face in the image. The output of the system is the position of subcitra containing faces were detected.
3. LIMITATION OF PROBLEMS
In the face detection system is given restrictions the following issue:
· Image inputs used are black and white with 256 levels of gray (grayscale).
· Face to be detected is the face facing the front (frontal), in an upright position, and no partially obstructed by other objects.
· The method used is a neural network multi-layer perceptron with back-propagation training algorithm.
4. RESEARCH OBJECTIVES
The research aims to create a design and implementation of a face detection system with the input of any digital image. This system will generate subcitra containing faces were detected.
5. BENEFITS RESEARCH
The results of this research can be used as a first step to establish a comprehensive system of face processing, which can be applied to face recognition or verification system faces. The application program can also be made of materials for further research in related fields.
With certain adjustments, the methods used may also be used for general object detection system that is not just limited to the face, such as vehicle detection, pedestrian, production materials, and so on.
From the results of this study are also expected to obtain a better understanding of the neural networks and the influence of various parameters used for the performance of the neural network classifiers.
6. METHODS
The method used in this study consists of the following steps:
· Conducting research literature on various references relating to research conducted. Topics to be studied include: pattern recognition, digital image processing, object detection in general, face detection, and artificial neural networks.
· Setting up a training data set that will be used for the learning process of the system. The data used will face through praproses be a 20x20 pixel resizing, masking and histogram equalization.
· Designing a face detection system with artificial neural network, then make the application program.
· Conduct training on the system with the training data set that has been prepared beforehand.
· Conduct performance testing system. The performance of the face detection system is measured by calculating the detection rate and false positive rate.
7. RESEARCH SCHEDULE
No.
Activity
Month / year
October
03
Nop
03
December
03
January
04th
February
04th
March
04th
1
Bibliographical Studies
2
Proposal Writing
3
Data Collection
4
Making System / Program
5
Testing System
6
Final Report
8. REFERENCES
ON DIGITAL IMAGE
(Thesis Proposal)
presented by
NamaMhs
NIM: XX.YY.ZZZ
To
INFORMATICS ENGINEERING DEPARTMENT
STMIK STIKOM BALIKPAPAN
APPROVAL SHEET
Thesis Proposal by title
DETECTION SYSTEM
HUMAN FACE ON DIGITAL IMAGE
presented by
NamaMhs
NIM: XX.YY.ZZZ
been approved by the Department of Information STIKOM Balikpapan with lecturers:
1. ......................................................................
2. ......................................................................
Balikpapan, dated ........................
Chairman of the Department of Information
Setyo Nugroho, ST, MKom
DETECTION SYSTEM IMPLEMENTATION
HUMAN FACE ON DIGITAL IMAGE
1. BACKGROUND
Nowadays more and more facial recognition technology is applied, among others, for biometric recognition system (which can also be combined with other biometric features such as fingerprints and voice), search and indexing system on a database of digital images and digital video databases, security systems access control area limited, video conferencing, and human interaction with computers.
In the field of face processing research (face processing), the detection of human faces (face detection) is one of the very important initial step in the process of facial recognition (face recognition). Face recognition system is used to compare an input face images with a face database and generate a face that best matches the image if any. While the face authentication (face authentication) is used to verify the authenticity / similarity of a face with a face data that has been entered previously. Areas of research that are also associated with face processing is localized face (face localization) is face detection, but with the assumption that there is only one face in the image, the face tracking (face tracking) to estimate the location of a face in a video in real time, and facial expression recognition (facial expression recognition) to recognize human emotions (Yang, 2002).
In certain cases such as photo shoots for ID card, driver's license, and credit cards, the images obtained generally contains only one face and have a uniform background and lighting conditions prearranged so that the face detection can be done more easily. But in other cases often obtained images containing more than one face, has a varied background, which is not necessarily lighting conditions and faces varying size in the image. An example is the image obtained at the airport, terminal, building entrances, and shopping centers. In addition to images obtained from photographs in the media or video recordings. In the general case of faces in the image has a shape very varied backgrounds.
This study will focus on the problem of face detection. With face detection system that is accurate, then the next process that face recognition can be done more easily.
2. PROBLEM FORMULATION
Face detection problem can be formulated as follows: the input is an arbitrary digital image, the system will detect if there are human faces in the image, and if there is then the system will tell how many faces are found and where is the location of the face in the image. The output of the system is the position of subcitra containing faces were detected.
3. LIMITATION OF PROBLEMS
In the face detection system is given restrictions the following issue:
· Image inputs used are black and white with 256 levels of gray (grayscale).
· Face to be detected is the face facing the front (frontal), in an upright position, and no partially obstructed by other objects.
· The method used is a neural network multi-layer perceptron with back-propagation training algorithm.
4. RESEARCH OBJECTIVES
The research aims to create a design and implementation of a face detection system with the input of any digital image. This system will generate subcitra containing faces were detected.
5. BENEFITS RESEARCH
The results of this research can be used as a first step to establish a comprehensive system of face processing, which can be applied to face recognition or verification system faces. The application program can also be made of materials for further research in related fields.
With certain adjustments, the methods used may also be used for general object detection system that is not just limited to the face, such as vehicle detection, pedestrian, production materials, and so on.
From the results of this study are also expected to obtain a better understanding of the neural networks and the influence of various parameters used for the performance of the neural network classifiers.
6. METHODS
The method used in this study consists of the following steps:
· Conducting research literature on various references relating to research conducted. Topics to be studied include: pattern recognition, digital image processing, object detection in general, face detection, and artificial neural networks.
· Setting up a training data set that will be used for the learning process of the system. The data used will face through praproses be a 20x20 pixel resizing, masking and histogram equalization.
· Designing a face detection system with artificial neural network, then make the application program.
· Conduct training on the system with the training data set that has been prepared beforehand.
· Conduct performance testing system. The performance of the face detection system is measured by calculating the detection rate and false positive rate.
7. RESEARCH SCHEDULE
No.
Activity
Month / year
October
03
Nop
03
December
03
January
04th
February
04th
March
04th
1
Bibliographical Studies
2
Proposal Writing
3
Data Collection
4
Making System / Program
5
Testing System
6
Final Report
8. REFERENCES
0 Comments
