hello, just a friendly beginner. i am planning to create an Arduino device that can track your eyes and eye gestures like blinking to control the mouse cursor movement and its functions like right click and left click but sadly i cant find any reference for the list of materials and the procedure on the internet. your help will be greatly appreciated😊
Seeing that you did not write a tutorial, your topic has been moved to a more suitable location on the forum. Introductory Tutorials is not for questions.
google is your friend
https://www.google.de/search?as_q=arduino+Eye+tracking
I have some doubts that the calculation-power of an arduino is big enough to realise Eye-tracking
The Arduino microcontrollers specs are
RAM 0,000002 GB
Flash-memory 0,000032 GB
direct bit operations 8 bit
CPU-Clock: 0,016 GHz
compared to a modern PC
RAM 8 GB
HDD-memory 256 GB
direct bit operations 64 bit
CPU-Clock: 2 GHz
or in smaller units
The Arduino microcontrollers specs are
RAM 2 kB
Flash-memory 32 kB
direct bit operations 8 bit
CPU-Clock: 16 MHz
compared to a modern PC
RAM 8.000.000 kB
HDD-memory 256.000.000 kB
direct bit operations 64 bit
CPU-Clock: 2000 MHz
1 Like
Yeah, that's because it's beyond the capabilities of Arduino. Maybe with some of the latest Arduino compatible boards like ESP32-cam it might be possible, but even if that's true, such a project would be way, way above beginner level.
Another possibility which might be more achievable would be to use an accelerometer sensor connected to an Arduino that has HID capability, such as a Pro Micro (don't confuse with Pro Mini). This could be used to monitor head movements to move a mouse cursor.
1 Like
hi thanks for your reply! is the concept within the capabilities of the raspberry pi?
Perhaps. Especially now that high quality camera modules with different lenses are available for Pi. But still the project would be at a highly advanced level. If you could make it work, you should be awarded a PhD!
1 Like
it depends in which way you want to detect the Eyeball-position.
The Arduino-based project used an array-line of infrared-reflection sensors which requires to wear glasses with the sensors mounted next to the eyes.
Though some commenters posted that it did not work for them.
If you want to use image-processing then an arduino is way too small
I guess it will be doable on a raspberry pi or a banana-pi
But as @PaulRB already said.
This is not a beginner-project. If you start learning programming from scratch
6 month = 20 days/month * 6 hours / day = 720 hours learning to write code
1 month 20 days * 6 hours / day = 120 hours to learn to use OpenCV
1 month 20 days * 6 hours / day = 120 hours to learn how to configure a raspberry Pi to be able to use OpenCV
3-6 month = 20 days/month * 6 hours / day = 360 to 720 hours developping your system to work reliably
maybe I'm too pessimistic. Maybe it can be done in 100 hours using a Raspi or a PC and using pre-configured OpenCV-configurations (where somebody else has put 2000 hours of work into it)
Is this just your personal interest or some kind of student has to realise some informatic project?
best regards Stefan
1 Like
hi stefan! this is a school research project. my initial plan was to use arduino and use led sensors next to the eye with a dismantled webcam to track eye movement. Unfortunately, the project i saw was only limited to eye tracking and can’t control cursor movement just like how I intended.
As soon as you have managed to retrieve the movement-data
moving a mouse-cursor is a piece of cake
The hard part is to reliably detect the eye-movement.
Your picture how such a project "works" seems to be
let's find the parts
let's find the code
cobble it together and you are done
This shows that you are greatly underestimating the time it takes to build such a system from scratch. And that you are greatly underestimating the know-how that is required to make it work.
Coding-knowledge
Opto-physical-knowledge
electronic knowledge
It is 1.000 - I repeat a thousand times harder to make this work than to search for an exotic iPhone-App in the Appstore that can do XY
Searching this App is a 20 hours job to find that one App than can do XY.
After these 20 hours you are done.
Your eye-movement-project has 500 researchings of the same size as the singular one for the app.
Using a raspberry-Pi
installing any Raspbian operatin system OK.
recognising oh! the OpenCV-software works only with this OS-version called XY
where can I donwload it? another 20 hours of research
downloading it.
trying to install OpenCV. Oh! my OpenCV-version is not compatible with this HDMI-driver!
looking up a raspberry-pi user-forum after 2 days they say "ask in this OpenCV-forum"
another 5 days to just get a cryptic answer full of technical terms you don't know yet
"install the ABC-package version 34.5 and all its dependencies"
How do you install the ABC-package?
What is the process of using GitHub?
What is the process of using Git?
How do I have to vary the sudo-command for my Raspian-OS to get this dammed package installed?
How do I get super-user-rights on my raspberry pi?
and these steps are only prelimary to start configuring OpenCV!
Boy If you have some other ideas what to realise as your school-project post these other ideas
that more experienced users than you can give advice what kind of project has the right size for the amount of hours you can spend on this project.
Estimating the time it takes to finish a project is one of the most needed basic skills of any engineer.
1 Like
i really appreciate your effort stefan! thank you for a much needed eye opening like this for a beginner like me.
Ok. Do you have any kind of specification or requirements that your school-project must fulfill?
is "do something with programming" already enough?
To quote a real nonsense-project I guess if your project-description would be
"I'm hanging in at home on the couch writing down every 10 minutes how I feel or what noises I hear"
this would be too less effort for a research-project.
There are a lot of things possible that are a lot of easier than recognising eyemovements where you can do research.
Research in the sense of
- building a device that measures something (temperature, humidity, air-quality, light, noise,
- counting objects or people passing by
If you are willing to spend some money:
- heartbeat,
- ppm CO2 in the air of a closed room
they are all much easier because there are sensors ready to use with example-codes
and interfacing a sensor with a digital interface requires learning about electronics a few hours but not months.
As your first idea is mouse-movements invoked through eyemovments
using a 3-axis G-force 3-axes gyro to invoke mouse-movements through turning the head
could be a research-project to examine how small can the head-moves be to beeing still able to control a mouse-arrow by pitching / yawing the head or a hand.
In the sense of starting research and obtain as the result
how much knowledge and / or specialised equipment is needed to make such a glove like in the video work
I mean if the final result is: how far can you come in the development process if you start from zero trying to develop such a glove as shown in the video below used by imogen heap
The final research-result is: on this journey towards the final working device by investing X hours of time you can reach ....
This might be possible too.
Or as a variant: how much can you do with a device that uses only one bend-sensor?
best regards Stefan
P.S. I'm pretty sure that in 100 years the original recording of imogen heap hide and seek
will be used as a famous example of very early electronic music using harmoniser
1 Like
thank you very much stefan! I have learned so much from this encounter. This is actually for our capstone research and the expected timeline of our teacher is to start building the device from last week of november until june
Hi azukano,
this is nice to read.
I googled for "what is the meaning of capstone-project"
and found this link
Aha !
Quoting
Capstone projects are generally designed to encourage students to think critically, solve challenging problems, and develop skills such as oral communication, public speaking, research skills, media literacy, teamwork, planning, self-sufficiency, or goal setting—i.e., skills that will help prepare them for college, modern careers, and adult life.
hm well this description and yout time-line
makes me think new
hm - what do you think how many hours per week is realistic to work on this project?
Do you have to do this project as a single person or are you allowed to work in collaboration with other students or even with other people?
Maybe collaborate with other people still fits into the 21st century skills like cooperation and communication.
This project is a great role model for serving the common good.
And as your initial idea is a device for tracking eye-movements for guiding a mouse-arrow on a screen.
Your idea of the eye-controlled mouse-cursor is doable but needs a big amount of man-hours and quite some knowledge in different fields to realise.
What is your idea where your device will be used?
There is one other consideration that your initial project failed to consider. Early on, perhaps 2002, my former company built circuit boards for a researcher attempting to use head/face movement to move a cursor and simulate a mouse click. He used a reflector on a person's fore head. Don't think he ever got it to work.
Your description completely ignores the head movement. That must be takes into consideration because the eyes may still be focused on a spot while the head moves and will affect any system that just "focuses" on the eyes.
1 Like
this is actually by group consists of 7 members. Realistically speaking, considering that we also have other subjects to pass like physics we can put around 15+ hours individually. If we are able to complete the eye controlled mouse-cursor, we are going to represent our school in a research competition in our country. We may have 7 members but the problem is we lack the basic know-how regarding electronics, robotics and codings. And our teacher requires us to build something with arduino or with other micro-controllers that are available to us (mostly only arduino microcontrollers are only available in our local shop and sometimes sell raspberry pi), a device that will help solve a certain problem in our society and in the case of the eye controlled mouse- cursor (if ever we are going to push through) our device is going to help people continue their work even after being amputated and aid them in using computers.
my initial plan was to recreate this for the eye tracking
but dont know how to make the device control the cursor
Which country do you live in?
It hsould be quite easy possible to get any kind of microcontroller
even it your country
i live in the philippines sir, i can order online but it would take weeks before it arrives and our teacher is already pestering us to already start the construction and there should already be progress by the first week of december
Well this is just a matter or organising a special transport at low-cost which delivers in a few days . Which would be a great add on to your project to organise it.
As I wrote earlier:
making a mouse-cursor move is a piece of cake when using a microcontroller with native USB
Take 30 minutes time and google for
Arduino mouse and similar keywords
post 10 links that you have found
do a first estimation from the 10 links you found which three ones do you think are most promising?
would you mind share the region or town next to where you live in the philipines?
even me I know a philipine woman from earlier times I can ask
"do you happen to know somebody flying to the philipines next week?
Do you think this person can carry a small package with a present?"