Eyes Tracking Opencv python

[@AiPhile]

Change the color of RGB LED with eyes 👁️👁️, based on eyes Tracking project, here is link : kzbin.info/www/bejne/aYqnnXaIiJajm9U

[@kacemichakdi3048]

thank you. u realy did a great workk

[@informationDesigner]

Thanks, that's a great job Is it possible to calculate the angle of direction of the horizontal and vertical gaze of the iris?

[@AiPhile]

Here we are not tracking iris already mentioned in the tutorial, its future work

[@informationDesigner]

@@AiPhile Yes, I am expecting that soon I have a project on this and I have commented on estimating the angles of the iris Please help me in this regard

[@AiPhile]

Check out mediapipe python module, you will get the iris tracking but it would be in c++, currently I am working on AR. Its take time to complete.

[@informationDesigner]

@@AiPhile Thanks for the advice

[@aswathyps4967]

Is it possible to access an external camera for eye detection than the system webcam

[@yuluscesar5822]

Very interesting work congratulation could I use it in raspberryPi4 thanks?

[@AiPhile]

Yeah, I have test that out it will work out of the box, just you to decrease Frame size that you get more Frame rate

[@pankhanafis818]

did it work for you?....was it fast enough? i am getting very low framerates....tried resizing the grayimage,but then it fails to detect faces

[@ManjeetSingh-tc8fs]

thanks a lot man.. u saved my life haha

[@AiPhile]

You are welcome,

[@kacemichakdi3048]

Hi sir, u realy did a grate job. I have one question , if we want to count the number of times that the eye see in the rigth and the number of time that the eye see in the left. thanks for this video.

[@AiPhile]

You can add counter of eah position of eyes., i would recommend you to check out my New video on this topic iris position estimation mediapipe: kzbin.info/aero/PLJ958Ls6nowVBQxl_-wmtN-c8xSBPJjPC

[@kacemichakdi3048]

Thank you a lot😀

[@sofizr1183]

Good 👍

[@AiPhile]

thank you!

[@Actrl51]

Could this be done dividing the eye in 9 boxes instead of three?

[@AiPhile]

Yeah its possible

[@aishwaryav3177]

hey, i need some help! firstly where did you run this code? which platform

[@AiPhile]

You will find the all the details on my GitHub repository, link in the description.

[@aishwaryav3177]

@@AiPhile i know..but where do i run it ? like pycharm, anaconda prompt

[@aishwaryav3177]

how can i connect with u? like mail id! got a couple of questions. please help it's important!

[@AiPhile]

You need installed on machine, then you need few python module following the instructions on GitHub repository,

[@aishwaryav3177]

I'm new to this domain

[@pankhanafis818]

good job bro....

[@AiPhile]

Thanks

[@pankhanafis818]

brother,i am trying to implement this in raspberry pi4 ,but the detection is too slow....i am getting 3.5 fps ....I have tried skipping every three frames in the main.py script while loop to get some extra frames.....I tried to resize the grayimage which is used as argument in the face detector method of the module script,but if i make it smaller,it seems like dlib doesnt look for faces in the whole image,and it skips most of the image parts,depending on how much i have made the frames smaller...can you please say the proper way of resizing images for dlib,so that it can detect images properly...thank you

[@AiPhile]

@@pankhanafis818 check this out kzbin.info/www/bejne/j17QdKehq5lkqM0

[@dyaa_unison5580]

could you use the tracked pupils and control the mouse with it, using autopy, pyinpt, pyautoGUI?

[@AiPhile]

You can checkout Iris tracking video tutorial

[@aswathyps4967]

We are doing a project based on eye detection for paralysed person which is introduced as specs(glass)... so that by detecting theri iris position the message is sent to the users phone whether the patient require food,water etc.Could you plz help us in this about what are the components that can we use and how to code this...

[@aswathyps4967]

Could you plz help us in this

[@SaimaKhan-gx8tj]

Hi.. Can this be done in google colab?

[@AiPhile]

Yeah, you have to modify the code bit and its difficult to access webcam on colab.

[@akcandlestick8948]

hi sir ji do you know about Python GUI Printing if we make some Billing GUI Programme Which Librrary we use for Direct printing on paper ??

[@AiPhile]

In future I will work on it, right now I am working on AR series.

[@akcandlestick8948]

@@AiPhile sir ji i am aware of this Printing library in Python if you know about pls reply msg

[@AiPhile]

I haven't used. This module at all

[@akashshetti3785]

There is a problem in installing dlib library using command prompt in windows 10 how to do it??

[@AiPhile]

I have mentioned in video tutorials all the reference.

[@akashshetti3785]

@@AiPhile while importing the packages like cv2 as cv it is showing error in visual studio

[@AiPhile]

Send screenshot on Instagram Please if possible

[@akashshetti3785]

@@AiPhile ok

[@akashshetti3785]

@@AiPhile ModuleNotFound error is coming for cv2

[@aswathyps4967]

Is there any sensor that can detect iris position

[@AiPhile]

Check out another video on channel about accurate iris tracking

[@Beast_Gaming_99]

I need your help right now...

[@AiPhile]

Yep, how

[@pal432-m5v]

47:07 How can you declare a function above its definition?

[@AiPhile]

?

[@pal432-m5v]

@@AiPhile You defined Position() later but the declaration came in the eye tracking function above. I am confused. Can u remove my confusion?

[@AiPhile]

All the functions are written in different file as module.

[@pal432-m5v]

@@AiPhile ok

[@kareemessam6245]

it gives me an error in the faceLandmarkDetector the error : File "C:\Users\Karim Essam\PycharmProjects\learnOpenCv\functions.py", line 74, in faceLandmakDetector landmarks = predictor(gray, face) TypeError: No registered converter was able to produce a C++ rvalue of type class dlib::rectangle from this Python object of type NoneType

[@AiPhile]

Send me screenshot, on Instagram or just check out updated version of same project on mediapipe here kzbin.info/aero/PLJ958Ls6nowUwRXHUcFwZy2CT0naMULR3

[@jyotiruhela9877]

I tried to run your project in my visual studio. but its giving error.i save the shape_predictor_68_face_landmarks file in predictor folder as u told me but its giving error.

[@AiPhile]

What kind of error can you show it to me?

[@jyotiruhela9877]

@@AiPhile Yes, import module as m and unable to open shape_predictor_68_face_landmark.dat 2 errors are showing in terminal

[@AiPhile]

@@jyotiruhela9877 make sure correct the path of Shape predictor.

[@jyotiruhela9877]

@@AiPhile I'd tried so many path. I extract the .dat file into predictor folder as u mentioned. Still its giving the error. Plz help

[@jyotiruhela9877]

Can u share me your email id. I'll share my screen on google meet if u want.

[@thunzard]

the predictor file is not opening sir, it's showing an runtime error. Please help.

[@AiPhile]

I have provided you link in GitHub readme, go over there and check out please, just download that,

[@thunzard]

@@AiPhile Sir I have downloaded the shape_predictor_68_face_landmarks.dat.bz2 file from the link provided in the Readme file of github, and I saved it in a folder called Predictor in the same directory as the code, but still when am running it it's showing the same error, and when am changing the file name from shape_predictor_68_face_landmarks.dat to shape_predictor_68_face_landmarks.dat.bz2 in the code, it is showing the version is not right it seems. I couldn't debug this error sir, please help.

[@AiPhile]

Extract the file first then you are good to go

[@thunzard]

@@AiPhile Sir can you tell the algorithm for detecting eye going top and bottom as well? Thank you so much for your help until now.

[@AiPhile]

Segment out the iris, using contours then you can track.

[@RajdeeppalSingh]

What if your camera is not so good and a bit burly

[@RajdeeppalSingh]

?

[@AiPhile]

Surely it will have effects, just give it shot, or try simpler one Eyes Tracking Mediapipe: kzbin.info/aero/PLJ958Ls6nowUwRXHUcFwZy2CT0naMULR3

[@RajdeeppalSingh]

@@AiPhile bro the board exams are starting from 15th November and it will be held online what do you think which software will the council use to check the eye position?

[@AiPhile]

Dm me Instagram we gonna talk about that.

[@osbornepereira7562]

It runs for a bit when I am checking the landmarks at 16:10 but I keep getting this error TypeError: __call__(): incompatible function arguments. The following argument types are supported: 1. (self: _dlib_pybind11.shape_predictor, image: array, box: _dlib_pybind11.rectangle) -> _dlib_pybind11.full_object_detection Invoked with: , array([[154, 154, 154, ..., 37, 34, 33], [153, 153, 151, ..., 40, 22, 25], [154, 151, 149, ..., 55, 27, 27], ..., [161, 168, 166, ..., 119, 121, 120], [164, 166, 166, ..., 120, 120, 119], [164, 164, 165, ..., 120, 119, 119]], dtype=uint8), None

[@AiPhile]

I would highly recommend you to check out the updated and more optimizated version of the project: iris position estimation mediapipe: kzbin.info/aero/PLJ958Ls6nowVBQxl_-wmtN-c8xSBPJjPC

[@osbornepereira7562]

@@AiPhile I am currently doing that, but do you know why this error happens? I have spent two days trying to understand this but I am unable to figure it out.

[@AiPhile]

@@osbornepereira7562 show the source code on Instagram, would let you know, but it seems you messed up with dilb landmark.

[@Notsonerdlife]

I am facing the same issue. Were u able to solve this?

[@AiPhile]

@@Notsonerdlife check out iris tracking playlist

[@frhadheydrov2546]

Exception has occurred: RuntimeError Unable to open Predictor/shape_predictor_68_face_landmarks.dat it give this error

[@AiPhile]

Down load the shqpe predictor link on GitHub repository