iitmcvg/eye-gaze

Repository for Eye Gaze Detection and Tracking

eye-gaze

Repositiory for Eye Gaze Detection and Tracking

We have implemented an Eye Gaze tracking system ( currently in Beta version ) using a series of algorithms. It computes the following :

• Pupil detection

• Facial Normal

• Gaze direction

To have a look into how we did it, just clone the repository and checkout v1.0.

$ cd eye-gaze
$ git checkout tags/v1.0
$ make

Or you can download eye-gaze v1.0 directly from releases.

For a trial,

$ cd eye-gaze
$ ./bin/oic

Dependencies

• OpenCV ( used 2.4.9 )

Sample outputs

References

Head-pose estimation

1.Michael Sapienza and Kenneth P. Camilleri - “Fasthpe: A recipe for quick head pose estimation”

2.Michael Sapienza - “Head Motion Tracking and Pose Estimation in the Six Degrees of Freedom”

3.Roberto Valenti, Nicu Sebe , and Theo Gevers - “Combining Head Pose and Eye Location Information for Gaze Estimation”

4.Xuehan Xiong ,Fernando De la Torre - “Supervised Descent Method and its Applications to Face Alignment”

5.Hiyam Hatem , Zou Beiji , Raed Majeed , Jumana Waleed Mohammed Lutf - “Head Pose Estimation Based On Detecting Facial Features “

6.Oliver Jesorsky, Klaus J. Kirchberg, and Robert W. Frischholz - “Robust Face Detection”

Pupil localisation

1.Fabian Timm and Erhardt Barth - “ACCURATE EYE CENTRE LOCALISATION BY MEANS OF GRADIENTS”

2.Tom Heyman , Vincent Spruyt , Alessandro Ledda - 3D Face “Tracking and Gaze Estimation Using a Monocular Camera”

3.Luke Allen and Adam Jensen - “Webcam-based Gaze Estimation”

4.Jian-Gang Wang , Eric Sung , Ronda Venkateswarlu - ”Eye Gaze Estimation from a Single Image of One Eye”

eye-gaze's Issues

Get X and Y Coordinates of Eye

Hi,
Thanks for your great work , i want to get X and Y coordinates in order to use them in another Project , still can't figure out the Code yet and can't find the Correct Variables if you are using Ones .
Thanks .

how to computer gaze in the code

hello, would you tell me how to compute gaze and which paper can i find the theory ?

color.cpp issue

Hi, I have OpenCV 2.4.9 installed and I am able to compile the program but when I ran it
./bin/oic
I get the following error:

OpenCV Error: Assertion failed (sin == 3 || sun == 4) in cvtColor, file /home/deeplearning/opencv-2.4.9/modules/imgproc/src/color.cpp

Do you know what the problem could be?

Thanks in advice!

can't find tags/v1.0 anywhere

when I run git checkout tags/v1.0, there is no such file tags/v1.0

what is the difference between master and demo?

Also - is there a cmakelists.txt.
I'm trying to compile this on macOS, i have a hard time getting x11 to link.

I wonder why you used X11,Glut and all that, i think all of this could have been done within opencv?

Thanks

computation of yaw angle

Hello iitmcvg,
I think you have make a wrong definition on yaw angle, the origin in code was
yaw = acos((abs(normal[2]))/(sqrt(normal[0]*normal[0] + normal[1]*normal[1] + normal[2]*normal[2])));
It should be acos((abs(normal[2]))/(sqrt(normal[0]*normal[0] + normal[2]*normal[2])));
or
acos((sqrt(normal[1]*normal[1] + normal[2]*normal[2]))/(sqrt(normal[0]*normal[0] + normal[1]*normal[1] + normal[2]*normal[2]))); right?

"Illegal instruction (core dumped)" error on Ubuntu 16.04 32 bit

Hi,
First of all congratulations for making such a project. I cloned your repository and did make clean before running make compilation was successful and I went to bin directory and ran ./oic but I got Illegal instruction (core dumped) error. I don't know what the error is. Only this binary runs :

  ./facegesmatch 
  DTW Score : 0

the rest crash.
I hope you help me.
Regards.
Ken.

about some variables in code

hi, thank you for sharing code! I have some questions about some varibales in code, I hope you can explain them for me, thank you again!

1. normal in FacePose, does it mean normal vector of 3D face plane
   2._CP, _LR, _CR, _CM, CE, like as vec_cp_kalman_l, vec_cp_pos_l

Could the project run on windows system?

Could the project run on windows system?

about the CMakelist.txt

I want to run the program on Win7 with VS,but I do not know how to do that without CMakelist.txt.
Could you do me a favor？
Thank you so much.

X-Y coordinates for Mouse Pointer movement

Hi,
I really appreciate your work, congrats.
I'm studying on it for a serious project, but I'm having a simple issue: I don't find the exact variables to use in order to move my mouse pointer on the screen, by looking to the webcam.
Could you help me, please?
(I'd like to use xdotool in order to move the mouse pointer, on Ubuntu 17.04, if this is an useful info for you).
Thanks a lot anticipately.
Regards.

Any plans to port the library to OpenCV 4.x?

Hi,
Any plans upgrade to the latest OpenCV and dlib versions?
Thank you.

Run on custom video/image?

Is this possible?

If yes, please let us know how to do it - or what script to modify to make it work

error occur in kalmanFilters.cpp

in the row 73 of the kalmanFilters.cpp. the size of the mat is 4x4, but the code show 17 numbers. Please what is the correct mat ?
KF_e_l.transitionMatrix = *(cv::Mat_(4,4) << 1,0,1,0, 0,1,0,1,0, 0,0,1,0, 0,0,0,1);

This program cannot run

Can the program run ? When I compile, there are a lot of code errors. Can you give me a program that can be used normally? Thinks a lot !

Help me map eye gaze on screen

Hi,
I would like to know if there is a way to obtain the values of the coordinates about on where a person looks at the screen.
If there is nothing already done for that problem, do you know how to do it or which information could I use and how to use them?
Thank you in advance for your help.

Error LNK2001 eye_gaze v1.0

@Aravind-Shankar Hi, i try to run this program but i keep getting those errors (you can see them in the attachment Debug_Errors). I'm using Visual Studio 2012, opencv 2.4.9, this is how i include the dlib file
C:...\Downloads\eye-gaze-1.0\eye-gaze-1.0\src

Debug_Errors.txt
Thank you for your help.

Issues related to pose estimation head

Hi,
I want to ask a question about pose estimation head. Does this program calculates the values of pitch, yaw, and roll? If so, how do you get these three values ? I got these three values through the code.
(std::cout << "Yaw : " << face_pose->yaw_180.0 / 3.14 << std::endl;
std::cout << " Pitch : " << face_pose->pitch_180.0 / 3.14 <<std::endl;
std::cout<< " Roll : " << face_pose->symm_x*180.0 / 3.14 << std::endl;)

But I found that the value of roll is greater than 5000 .It is so strange ! Could you give me a suggestion ?