Open enrico310786 opened 2 years ago
Hello @enrico310786,
in this code, I have set a virtual screen just below the camera without any rotation. I recommend calibrating the position of the screen as described by Takahashiet et al., which provide an OpenCV and MATLAB implementation, for a better accuracy. In the paper you mention, the authors call this approach "classical geometry-based mode". They also propose using a "mirror-based calibration technique" but they use the method discussed by Rodrigues et al..
Hi, I case of calibrating the position of the screen. How it is used in your code?. Is a modification of the method "plane_equation" required in case of rotation between the screen and the camera?
TODO load calibrated screen position plane = plane_equation(np.eye(3), np.asarray([[0], [0], [0]])) plane_w = plane[0:3] plane_b = plane[3]
Thanks
@enrico310786 Did you manage to figure this out? I'm having the same issue
Hi, I cannot understand how the calibration data obtained with the main.py script in the gaze-data-collection project is used in this project. In that case a .csv file is produced and the calibration images are saved. How and where are this data used to optimize the gaze vector projection from the 3D space to the point on the monitor on the 2D space? The paper "Efficiency in Real-time Webcam Gaze Tracking" talks about three ways to perform this type of monitor calibration. Geometric, machine learning and hybrid? What kind of optimization do you apply?
Thanks