The module learning outcomes addressed by this assignment are:
Evaluate current research topics relevant to intelligent systems for practitioners, professional/legal bodies and industry stakeholders;
Undertake a significant piece of research in a topic currently relevant to intelligent systems subject area;
Effectively communicate research findings in the form of a research paper to a non-specialist audience.
There are two assignments for KF5042 accounting for 100% of the marks for the module.
This assignment accounts for 60% of the overall marks for the module.
Any queries should be directed to:
Face recognition involves comparing an image with a database of stored faces in order to identify the individual in that input image. To achieve this, the images can be analysed to determine distinguishing features so that faces can be identified and recognised. There exist various methods of face recognition involving a series of steps to capture (image acquisition phase), extract the distinguishing face features (feature extraction) and compare the extracted features (classification) of a face to a database of stored images.
Your task is concerned with a comparative study of face recognition between Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA).
PCA based face recognition was covered in the lectures and in the Lab sessions.
You are required to undertake a critical piece of research on face recognition. By revisiting face recognition using Principal Component Analysis (PCA), also called eigenface based recognition, you are required to critically analyse and evaluate the performance of the LDA method including a comparative study against its PCA counterpart. Performance in terms of recognition accuracy rates and false recognition rates should be considered and critically discussed using the database provided.
The report should describe face recognition technology including:
a detailed description of the two algorithms (PCA and LDA) for feature extraction
a comparative study of the discriminative power of each method
a description of the classifier and the distance metric used
a description of the learning and testing phases when implementing the recognition system
a description of the generation of the results and their analysis
This is an individual report. You will submit your findings in the form of a report of no more than 1200 words for submission by the deadline specified.
Important Note: It is the responsibility of each student to ensure that his/her report is submitted by the deadline.
You are required to write a research paper to critically analyse and discuss the performances and significance of PCA and LDA when applied to face recognition.
Your paper should consider the following guidance and contain the following subtitles as a minimum:
Title page (includes title and student name and student number)
An introduction using appropriate information and problem statement of face recognition and its applications
The significance and how face recognition works defined alongside details of how the paper is organised.
Scope: describe the various stages of a face recognition system; give a diagram of the various stages and clearly describe the purpose of each stage
Feature extraction using PCA and LDA methods
How each method extracts the features from face images
Discuss the issue of feature dimensionality problem and how it can be solved
Clearly compare and contrast the discriminative power of each method
Scope: Description of the feature extraction of each method (e.g., how the features are extracted; how feature vectors are generated for each class/face; how the resulting large feature vectors are reduced while still keeping the discriminative power intact.
Describe the (training) learning and testing phases? How many face images are used for training (learning) and testing? What is the effect of the number of images used. For each experiment, clearly mention the number of images used for both training and testing phases.
Describe the tools used for the classification (distance metric and the classifier used)
Results and Analysis
Using the YALE Face Dataset you are required to revisit the Lab on PCA based face recognition. Run the face recognition application using the dataset by splitting the dataset into 20% for learning and 80% for testing. Using all the image classes of the dataset generate the recognition accuracy and misclassification rates. Repeat the process by varying the split to 30%-80% and then 40%-60% and record the performance as above.
Now create face recognition system using LDA (i.e., replace the PCA algorithm with LDA) and repeat the above experiments using the same dataset split rates. Record the performance findings.
Compare the performances obtained and give a critical discussion, reflection and analysis of the two methods.
Use Tables and Figures to show the results
You should give
Clear evidence of excellent critical reviewing and analysis of the performances and significance of the two face recognition methods
Scope: use Tables and Figures to reflect and analyse the performance; discuss their significance by giving a comparative study of the two methods (PCA and LDA) in relation to face recognition, include some screen shots of the results.
A brief summary of the key findings established from your research
A full list of references used within the paper should be provided. The Harvard Style of referencing should be applied throughout the assignment.
MATLAB code of the applications must be attached
Please adhere to the following requirements:
Submission will be via Turnitin on Blackboard; please see the front cover for the submission date. Please note, you can submit your report formatively to check for originality i.e. to help check for potential academic misconduct in the form of plagiarism. You can do this multiple times. However, you must ensure the last attempt is your final summatively assessed paper and is correctly submitted ahead of the deadline indicated on the cover page.
References should be excluded from Turnitin checking
The report should be written in a formal reporting style and without use of personal pronouns (for example, no use of ‘I, me, my, our, we, they, he, she’). If you find it difficult, you may want to research the use of the passive voice; help is also available via skills resources online (see above).
Word count 1200 excluding Matlab code, references and Appendix(ces).
Only Microsoft Word or PDF file formats will be accepted.
Format: Use IEEE 2-column conference format (see the template: https://www.ieee-pes.org/images/files/pdf/pg4-sample-conference-paper.pdf)
Layout should make reasonable use of margins, clear headings, single line spacing and font size should be 11pt (i.e. your report should be professionally presented).
Include page numbering, words count, the module code and your student ID.
All content including references and appendices (if used) should be contained in a single document. Title, author’s names, Appendices (if used) and References are excluded from word count.
Referencing should be in the Harvard style (https://libweb.anglia.ac.uk/referencing/harvard.htm
You must adhere to the university regulations on academic conduct. Formal inquiry proceedings will be instigated if there is any suspicion of plagiarism or any other form of misconduct in your work. Refer to the University’s Assessment Regulations for Awards if you are unclear as to the meaning of these terms. The latest copy is available on the University website
Where the report includes someone else’s words (quotations), they should be correctly quoted and referenced in accordance to the Harvard System. Help regarding referencing can be found at: http://www.durham.ac.uk/sd/central/library/resources/referencing/
DescriptionIn this final assignment, the students will demonstrate their ability to apply two ma
Path finding involves finding a path from A to B. Typically we want the path to have certain properties,such as being the shortest or to avoid going t
Develop a program to emulate a purchase transaction at a retail store. Thisprogram will have two classes, a LineItem class and a Transaction class. Th
1 Project 1 Introduction - the SeaPort Project series For this set of projects for the course, we wish to simulate some of the aspects of a number of
1 Project 2 Introduction - the SeaPort Project series For this set of projects for the course, we wish to simulate some of the aspects of a number of