Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
This is an example of the kind of writeup the OCR Computer Science NEA project needs in order to get an A*
Typology: High school final essays
1 / 231
This page cannot be seen from the preview
Don't miss anything!
Candidate Name: Candidate Number:
Page 1 of 231
H446-03 – P CONTENTS
Candidate Name: Candidate Number:
AddNewClass.py 220
Registra on.py 225
The Problem
At the moment teachers have to manually register students. A student's name is read aloud and the student responds, iden fying if they are present within the lesson. There are many problems with this current model of taking the register. One being that there is a responsibility for the teacher to remember to take the register at the beginning of each and every lesson. By a emp ng to complete this task, nearly 10% of the lesson is spent reading out the names of each student and the students having to wait and listen for their name to be read out, risking a name being missed.
The register is only taken once at one point in me and if someone is late, the register has to be manually updated again and I’ve observed the frustra ons of teachers where precious me is wasted as me is spent talking to the student and marking them as late, impac ng the flow of the lesson as a whole.
To tackle these frustra ons, I thought it would be appropriate to create a school registra on system where pupils are registered via facial recogni on, instead of a teacher having to manually go through and mark pupils in. This system will work where the student will enter the classroom, log-in and open the registra on program where it captures an image of their face which can be used to iden fy and register the student.
To accompany this I’ll design a management system which is compa ble with the facial recogni on program. This will act as a place to handle registra on, misiden fica on of students and other problems which may arise.
This proposed solu on does not only help with reducing responsibili es for the teacher and saving me but abuse of the current system is prevented. In the Sixth Form you can sign yourself in and students will abuse this privilege by signing in their friends if they are running late, saving the late person from any punishments and retaining their punctuality percentages. By employing a computa onal approach, such as the use of facial recogni on, means that one student will not be able to sign in for another as your facial features is a biometric unique to you.
The Stakeholders ( Clients/ Users)
My client are the teachers, working at Secondary School and Sixth Form Centre. They administer a number of classes with each containing around 30 students. On top of the classes they teach, they may be asked to cover another teachers lesson if they are away. Each one of those less will require the register to be taken before the class commences.
Page 4 of 231
Candidate Name: Candidate Number: When I spoke to a client (Conversa on outlined on pg 4), they concerned that me is everything and the more students you have, the more me is lost, which is important for teachers to minimise as it greatly impacts the academic life of students as actual learning me is reduced.
The use of facial recogni on would mi gate any chance of a teacher forge ng to take the register and takes away the risk of the student not hearing their name which is pivotal for the safeguarding and wellbeing of students at the school.
I plan to conduct a formal interview with the client to gain a deeper understanding of any personal problems they find as well as ge ng an idea of certain features that the client desires.
Also, the end users will be not only the teacher, as they interact with the management students, but also the students will use a facial recogni on program I will develop in my first itera on. For this program, I plan to create mockup screens to gain feedback and guidance for the layout from the students.
Summary Conversation with a Stakeholder
Mr :
When I first introduced the idea of using facial recogni on to register students my stakeholder suggested to use fingerprints as the biometric for registra on, which would consist of having a physical hardware component to use op cal imaging to allow for a more accurate iden fica on of the student. Moreover, if you take a conven onal op cal fingerprint sensor then at any me of day with whatever ligh ng, it will be able to gain an image of the print.
However, I argued that fingerprin ng, if there is one sensor, will require students to queue up to register. Each student will scan their finger one by one hence it is almost as me-consuming as a teacher manually doing the register but with facial recogni on each student can walk into the classroom and scan their faces seamlessly, taking a frac on of the me with the cost of accuracy.
Furthermore, since the facial recogni on system will be implemented in a classroom environment, each room will be well lit hence the ligh ng shouldn’t be an issue when detec ng faces.
The main point the stakeholder reiterated is that me is everything, anything to save me for the teachers is great.
JUSTIFICATION OF HOW TO SOLVE THE PROBLEM CAN BE SOLVED BY COMPUTATIONAL
METHODS
Thinking Abstractly
Page 5 of 231
Candidate Name: Candidate Number:
Decomposing Further
These categories will be split into their own individual scripts and further broken down into subprograms. This will help with code readability as well as isola ng any problems that are found quickly which is pivotal for a system that will be used every school day.
Thinking Concurrently
Page 7 of 231
Candidate Name: Candidate Number: INTERVIEW WITH STAKEHOLDER ANALYSIS(Teacher)
Interview Transcript can be found in the Appendices Page 194
For my first interview, I tried to gauge the impact that taking a register manually has and the addi onal problems that arise with the current system.
One key problem was the risk of forge ng to take the register. The client clearly outlined how it is a huge safeguarding risk and the school needs to know the whereabouts of each student. By employing the use of the proposed facial recogni on system there will be reduced human error where if teachers miss a students name or the student did not hear their name being called out, with the facial recogni on system the student will be registered as soon as they walk in, allevia ng the stress associated with remembering to take the register.
Another problem is that remembering supply teachers struggle to iden fy and register students so in my solu on a high chance of correct iden fica on is required. This is key as when there is a supply teacher a student will fake their name or pretend to be another student. With facial recogni on and iden fica on, you can correctly iden fy the student and prevent any of these games the students play with a new teacher from unfolding.
Furthermore, in the interview, it was stated that the register has to be taken within 5 minutes of the start of the lesson, which is hard to achieve if you are trying to se le down and manually take the register. The fully automated system will save me for the teacher as they will not have to go down a register, marking a student in one by one. The first itera on will be using the computer webcams where the students will manually take the pictures and if there is enough training data I can move it all to a raspberry pi where I can register students seamlessly as they walk in through a door saving even more me.
Page 8 of 231
Candidate Name: Candidate Number: SIMS provides analy cal data in the form of bar charts and in this case an a endance breakdown with green as present, red as authorised absence, yellow as unauthorised absence and grey as missing marks. Charts allow the teacher to get a quick visual summary of what the a endance is like for the class and school.
I’ve taken away that this is an effec ve way to form a summary as striking vibrant colours such as green, red and yellow are used. This means any sharp rise of absences will be easily no ceable rather than a teacher purely seeing numbers regarding a endance. As well as the teacher will be able to see clearly if anyone was missed and not marked in.
Toolbar
SIMS provides a toolbar which contains op ons such as taking the register, adding comments, finding future lessons, seeing class summaries and more. A toolbar like this allows a quick shortcut to features that will be most frequently used by teachers instead of them having to navigate through various drop down boxes. This will increase speed as well as it makes the User Experience more pleasant as the tools required are upfront and ready to use.
Timeline
Page 10 of 231
Candidate Name: Candidate Number:
A meline is provided showcasing the future lessons a teacher has at a given date and me. It is displayed in a grid format with the me on the Y-axis and the date on the X-axis. This allows teachers to have an overview of upcoming lessons. This enables the teacher to be no fied if a lesson is rescheduled or they are asked to take another class, making them quickly aware of any changes.
Page 11 of 231
Candidate Name: Candidate Number: up/down missing a student. Another feature is that SIMS instantly shi s to the next line when the code is generated, making it easier for the teacher to transi on to the next student.
Comments
SIMS gives the ability to provide a comment. This may be to give addi onal context regarding why a student is ill or late. Providing the comment and contextualising certain informa on allows all teachers to update and no fy each other of problems regarding a student in one central loca on.
Features I Didn’t Like
Multiple Drop Down Tabs To Take Register
I didn’t like how there were mul ple drop down menus un l you got to the part where you click to take the register. Having these numerous tabs makes it difficult for any new teachers to quickly grasp the system as ‘Take Register’ is hidden away. It would be be er if the register can pop up in the meline and a teacher can click/right click the class to take a register.
In my solu on, I will most likely use a le-based system so important func onality like registra on can easily be found.
Clicking ‘Save’ to Save Register
Page 13 of 231
Candidate Name: Candidate Number:
I don’t like how you have to click the ‘Save’ bu on for the register to save. This is due to the fact this means the teacher will have to remember to click the ‘Save’ bu on a er every register. An alterna ve solu on I’d suggest is to have live autosaves where every me a teacher updates the register, the update is saved.
In my solu on, the students will be registered as they walk into the classroom, so there is no need for a ‘Save’ bu on.
Description:
There are parallels between registering for an event and registra on for a classroom where the eventgoers are the students and the event staff valida ng their entry are the teachers. Zenus Biometrics is a provider of biometrics solu ons and facial recogni on so ware. The main use of their facial recogni on so ware is to cut down on registra on me at events. Instead of using a paper cket or an e- cket from your smartphone. The way in which this is achieved is that a photograph of the eventgoer is taken before the event as training data to be added to the facial recogni on so ware.
FEATURES APPLICABLE TO MY PROJECT
Using facial recogni on instead of a manned kiosk with an event staff checking each person’s cket is be er due to the fact the eventgoer doesn’t have to take out their cket, some may have not registered beforehand and the person manning the kiosk will take considerably more me scanning a QR code or typing in an event code to find out if they are a valid cket holder.
Reducing Check-in mes is important as it allows the eventgoer to spend more me at an event. This leads to be er enjoyment for them as well as a higher likelihood they will spend money on goods at the event or buying food and beverages along the way genera ng more revenue for the event planners. My client stressed how important saving me is and this would result in lesson me being saved which may lead to be er student grades.
Page 14 of 231
Candidate Name: Candidate Number: Asking an Eventgoer to Take a Headshot Prior to the Event:
The reason I didn’t like this feature is due to the fact it may seem somewhat inconvenient to ask an eventgoer to take a picture of themselves before going to the event. There could be issues with ligh ng, quality of the photographs which may affect if they will be recognised when they go to one of the kiosks at the event. Using facial recogni on may be faster than using an e- cket with a QR code however does asking an a endee to take a photograph waste more me for the eventgoer than it saves.
In my solu on, I will have to ask students to take a photograph of themselves which begs the ques on, will I save more me compared to a teacher manually taking the register. I believe that since this system will be used at least five mes every school day with over 750 students, the net me savings will outweigh the me it takes to capture images of students.
Not All Eventgoers Will Take a Headshot
In the case with Zenus Biometrics, a endees opted in to use the facial recogni on system. This indicates that not all people wanted to use the system. This may be since people didn’t feel comfortable with a company holding an image of their face or many just forgot to use the system. In this scenario, that means you will have a mixture of people using e- ckets as well as using facial recogni on which may lead to irregulari es in the me it takes to register an eventgoer as some will be using e- ckets and others facial recogni on.
In my solu on, as I’ve stated above, I will get the image of the student; however, all students will be using the facial recogni on system. If the solu on was on an opt-in basis, then it will probably take even more me to register the students with the facial recogni on system and to do the register normally.
A er breaking down the problem, speaking to stakeholders and looking at relevant features in exis ng solu ons I’ve iden fied some of the key features to be:
● Tile Based Format : As outlined, this will facilitate teachers to clearly see what key func ons are available to them making it to new/supply teachers using this system. ● Send a message to attendance officer if a student is 10 minutes late : The client emphasised how this is a feature that they would like. The feature is necessary to alert staff of a missing student as the safety of students is one of the schools top priority. ● A teacher won’t have to take register: Teachers will no longer have to take the register manually as all the registering is done automa cally by the automated system. This is achieved by the facial recogni on system iden fying the student and that data being sent to the central system management database server to mark a user as in.
Page 16 of 231
Candidate Name: Candidate Number: ● Charts : Charts are necessary for the visualisa on of data. This will let teachers easily iden fy if a student has a track record of being late as the chart system can plot correla on and trend lines or quickly iden fy if any student is missing. ● Codes : Codes allows for the explana on of the absence of a student to be summarised by a single le er. This means by a single le er, e.g I, the teacher can know the student is ill and prepare any work for the student to catch up with if necessary. ● Comments : Comments allow for addi onal informa on to be added which codes may not be able to provide. An example of this may be that if a student is late, the reason why the student is late can be known and measures can be put in place to help improve the students a endance. ● Log-in System : A login system is necessary to prevent unauthorised people from having access to the database. This is necessary to be in compliance with GDPR (Integrity and confiden ality:h ps://ico.org.uk/for-organisa ons/guide-to-data-protec on/guide-to-the-gener al-data-protec on-regula on-gdpr/principles/) as I need to handle personal data in a way I protect the iden ty of my client. A way of achieving this is by encryp ng the images I am going to keep of students so anyone with access to the database can’t go through a data users informa on for no reason as well as if the database is compromised, an unauthorised person will not be able to see the images.
LIMITATIONS
- Misidentification of A Student A problem that could occur is that a student could be misiden fied leading to the registra on of the wrong student. This could be because the student could have a twin hence having very similar facial features or it could be due to the ligh ng, angle or environment of the camera shot, making it difficult to match it with a photograph that was shot in a different se ng.
To combat this there will be a misiden fica on system where the student can manually correct any errors due to the system. If the student is logged into the computer, I can check the username of the account and figure out who the student is without their input. An indica on that a student has been misiden fied is if a student from a different class is iden fied or a student is marked in twice. A teacher can be sent a no fica on of this misiden fica on and the image of the student.
- Security Problems With Holding Faces Of Students Since I am holding other students photographs it is my obliga on, legally, to store the photographs in a way where access to the photographs is restricted and safe as well as if the photographs were compromised, they are in an unreadable form. Prior to me gaining the photographs for training purposes I would need to find out which students would like to par cipate with tes ng of my registra on system. Following that, I would have to give them a rundown of how their data will be used and I will have to be ready to erase or delete any data if requested.
To ensure physical security, the raspberry pi containing the database will stay within school grounds, preven ng anyone outside the school ge ng their hands on or tampering with the raspberry pi and its contents.
Page 17 of 231
Candidate Name: Candidate Number: OpenCV Library: OpenCV will be an essen al library for powering my facial recogni on so ware as this contains the CASCADES, RECOGNISERS, GRAYSCALE CONVERSION, VIDEO CAPTURE READ/WRITE/SHOW.
JSON Library: This is for wri ng and reading data that we have saved locally for the labels and training file.
Numpy Library: This is required for conver ng the pixels into a numpy array for training
Pillow Library: To easily open and manipulate the image, for example, conver ng it into greyscale.
Python 3.X.X: The code will be wri en in Python 3.X.X hence the program and a compa ble interpreter is required.
MySQL: This will be required for crea ng the structured, organised and persistent database.
phpMyAdmin: This will act as the database management system for my database with a useful feature of having a web interface allowing it to be accessed anywhere on the network.
tkinter: This is necessary for the Graphical User Interface for the School Management system.
mysql.connector: To establish a connec on to the database.
- Hardware
As python is able to run on something like a raspberry pi 1 it will not require any expensive new hardware for the program to run. Hence here are the requirements windows suggest for running Windows 7.
Minimum Requirements for Windows 7
1 gigahertz (GHz) or faster 32-bit(x86) or 64 bit (x64) processor
1 gigabyte (GB) RAM (32-bit)
16 GB Hard Disk Space
Monitor – For the teacher to interact with the management system and a ain a summary of class a endance. The student also requires a monitor to see if their face has been registered, iden fied correctly and that they have been marked in.
Mouse, Keyboard – Required to enter their username if in a training session or correct name if misiden fied.
Raspberry Pi – To host the database containing student images.
Webcam - This is required to actually take the image of students in the training sessions or when the facial recogni on so ware is running and actually iden fying the students. The best resolu on would be 480p as this will sufficient detail with a lower file size.
Page 19 of 231
Candidate Name: Candidate Number:
Requirements Justification Tests/Evidence
Confirma on box pop up once face is detected.
This is required as if anyone is misiden fied, the user can select no and manually enter their name to be manually marked in.
Proposed Solu on
Box around a detected face. This is needed so the user knows which face has been iden fied if there is a frame with mul ple faces.
Proposed Solu on
Camera input displayed on the screen
The input from the camera needs to be displayed to the user so they know if their face has been recognised and that they have been marked in.
Proposed Solu on
A ain at least 5 images of students.
These images are required so they can be used to train the recogniser.
Proposed Solu on
Send picture of face to teacher if student is misiden fied.
If a student is signed in twice or a student who is not in the class is signed in is registered.
Proposed Solu on
Encrypted file transfer of users images to database.
Keeping students images encrypted prevents someone from being able to view the images if they intercept the files during data transfer.
Proposed Solu on
Page 20 of 231
