Eye motion tracking - Opencv with Python 

🔥Learn how to build your own AI vision solutions: pysource.com/community

Thank you very much for the tutorial. I am curious what type of camera you used - Is it a web cam? What resolution and fps would you suggest for a setup. Thank You.

Thank you so much for all of your videos! I'm currently working with gaze detection at work, and you helped me a lot! :)

Thank you so much for this video as well as the eye gaze series. They are really awesome. I have faced a problem with the code in this video when obtaining the contours by the function findContours, it produces an error "not enough vales to unpack (expected 3, got 2)" I hope you can help me with that.

Thank you so much for the very interesting and helpful video. I have a question for you. What to do in order to count how many times eye iris moves for example in a minute? Thanks for consideration.

Hi, after detect where the eye is looking can I use this information and send it into the microcontroller to take some action? if yes, this code will be efficient to do this work? and thank you for sharing this awesome work.

Great video sir! Just one question, how do you calculate the exact position of the eye in opencv?

Its amazing that gpt3 can program its own code, so it can code this automatically, I was wondering if the cisco opendna spaces is what is used to route all the cloud data for the public.

Excellent and educative source, thanks, man!

Thank you, it’s extremely helpful.

Sir what is the hardware requirement? like what camera and all?

is there any algorithm in particular that you're using?

Thank you very much, your Tutorials help me a lot and I learn much even though i am just beginner. If you want, could you make a Tutorial about tracking objects like darts to find out their position in the dart board (wether it is a 20 or a 5 for example).

Hello, I am coding a project to where I have to set a variable to a different value when my eyes look in different directions. Whenever my eyes look away, I want it to make value = 'some number'. How could I do this?

When i try to run the contour part i got error like AttributeError: module 'cv2.cv2' has no attribute 'drawCounters' and ValueError: not enough values to unpack (expected 3, got 2)

Thanks a lot, your videos have helped me in my college work and since I am new to this, it really helped me. I just wanted to ask, if you could provide the codes to crop the eye part as you have showed.

is it possible to fiz the camera window os a HTML page? I mean nos a desktop app, but inside a web app?

What do I have to do if i want it to do this while i give an input video of my entire face

Hello sir from where i can download that eye movement video That is not in the discription!

Hi,Thank you for the video.Could you help me for tracking eyes of audience from live video conferencing like zoom meeting or meeting in MS team

Simple, amazing. I already learned a lot from different sources, but this videos comprises very well the main idea of ​​Computer Vision. I would just recommed add 2 things: Instead of using the '_' special variable, use a named variable, helps us that do not have much knowledge in cv2 to understand better what its been returning. Also, I would like to you to add a little explanation about each cv2 methods (Yes, I know that I can check the docs, but I want a shortcut :)

For some reason "_, contours, _ = " does not work for me. If I leave off the first '_,' then it does work, i.e., "contours, _ = " works. Same thing in the shapes video.

I got below error, TypeError Traceback (most recent call last) in 1 while True: 2 ret,frame = cap.read() ----> 3 roi = frame[269:795,537:1416] 4 gray_roi = cv2.cvtColor(roi,cv2.COLOR_BGR2GRAY) 5 gray_roi = cv2.GaussianBlur(gray_roi,(7,7),0) TypeError: 'NoneType' object is not subscriptable

how can i fix this error _,threshold = cv2.threshold(gray_roi, 3, 255, cv2.THRESH_BINARY_INV) IndentationError: unexpected indent

is it possible to use this as a mean to verify a person is real not a picture or some sort of that

how can we take this data of movement of pupil in csv file to perform ML Algorithm

why do you speficy position of the eye; roi = frame[269: 795, 537: 1416] rather than detecting eye first using custom object detection (with tensorflow) or already build haar cascade object detecion? With your approach it only applies to your recorded video witch have no sense. Imagine if you try to track eye motion with live webcam. Overall thank you for contributing

Hi... Will it work with VR device?

can we use erosion and dilation to eliminate the upper part where you don't want it in the pic? i find it easier than using blur and break the loop

what is Algorithm ? i need to know name Algorithm.

Hi .. I have a mistake is : Traceback (most recent call last): File "D:\eye_motion_tracking\tessst_roi.py", line 19, in contours = sorted(contours, key=lambda x: cv2.contourArea(x), reverse=True) TypeError: 'NoneType' object is not iterable what is solve?

Is it possible to create Tobii style application with webcam or better one?

Great tutorial sir, Can you give me direction how to do the same for live video example: using the camera. Thank you !

Hi buddy. What device did you use?

I'm getting error while executing code can u please help with it

but this method will not work in outdoor conditions. like in sunlight

How to detect liveness in a face

where could i run this source code ?

One of my questions is how would you make the console print out whether it is moving left, right, up, down, or center? You mentioned at the end of the video that you could do that. Could you please provide that information? Thank you!

I got another error at this code _, contours = cv2.findContours ( threshold , cv2.RETR_TREE , cv2.CHAIN_APPROX_SIMPLE)[-2:] for cnt in contours: # ctr = numpy.array(cnt).reshape((-1,1,2)).astype(numpy.int32) cv2.drawContours(roi,cnt , -1,(0,0,255) ,3) Error says, cv2.drawContours(roi,cnt , -1,(0,0,255) ,3) cv2.error: OpenCV(4.0.0) C:\projects\opencv-python\opencv\modules\imgproc\src\drawing.cpp:2509: error: (-215:Assertion failed) npoints > 0 in function 'cv::drawContours'

sir, can you give an idea about how we recognize iris and matched to our pre-stored iris image?

hi bro how i 4 led on off from eye movement ,

Great Work ! But can anyone help me to do this using Iris ??? What dimensions i have to change, please anyone guide me for this .

I have a question. When i try to run the code after the command cv2.drawContours(roi, [cnt], -1, (0,0,255), 3), i get the following error code: TypeError: contours is not a numpy array, neither a scalar

File "P:/Pysource/eye_motion_tracking/eye_motion_tracking.py", line 17, in _, contours, _ = cv2.findContours(threshold, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) not enough values to unpack (expected 3, got 2) this is the error, can anyone please help me?

How to calculate how much the change of position contour..?

What about tracking dilation

i need little help.how can i estimate the position like if i move my eye up then it just print eye location up.i want to use it for my project

I've got a problem in line 14 of your code. error: (-215:Assertion failed) npoints > 0 in function 'cv::drawContours' Can anyone figure out what this means?

Is it possible to output the coordinates of the center of the cross?

sir how can i detect the left, right direction ?? please help

I was waiting for you

i cant seem to find this face video you imported anywhere. Anyone know how I can get hold of it?

can we make it using the live video from a webcam??

i doubt this will work on other video with different light conditions...

_, contours, _ = cv2.findContours(threshold, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) ValueError: not enough values to unpack (expected 3, got 2) what wrong with that? Can someone help?

Thank you so much 😊

please help how to draw circles on top of thos contours instead rectangle

I think I've been hacked,my question is that if it's possible that this person is able to see what I'm exactly reading? Because I can hear them reading along with me. Sorry for my English

شکریہ

Hard coding the ROI is cheating. You want to train a HAAT network (there probably is one already) to detect the contrast levels that make up an eye and return that ROI.