search

hysts / pytorch_mpiigaze_demo

Gaze estimation using MPIIGaze and MPIIFaceGaze
MIT License
295 stars 66 forks source link

distortion coefficients: #3

Closed yongwuML closed 3 years ago

yongwuML commented 3 years ago

Thank you for the excellent work, I want to how can I get distortion coefficients, I find that is array [0,0,0,0,0,]. It is correct? I think it's not all zero.

hysts commented 3 years ago

Hi, @yongwuSHU

You can calibrate your camera using a checkerboard pattern and OpenCV functions. I think following articles would be helpful. https://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_calib3d/py_calibration/py_calibration.html https://www.learnopencv.com/camera-calibration-using-opencv/

yongwuML commented 3 years ago

hi, @hysts Thank your replay. If I got a video from internet, could I get the distortion coefficients by OpenCV functions? Sorry to bother you.

hysts commented 3 years ago

@yongwuSHU

I don't think it's possible to get the distortion coefficients in that case.

yongwuML commented 3 years ago

@hysts Thank you, actually, I want to pre-process the dataset EYEDIAP, it provided intrinsic parameters (martix 3*3), but no distortion coefficients. Could you give some ideas. Thank you.

hysts commented 3 years ago

@yongwuSHU

In my understanding, cameras in EYEDIAP datasets are already undistorted, but i'm not 100% sure, so you might want to contact with the authors.

yongwuML commented 3 years ago

@hysts Thank you, trully. I have sent some emails to authors , but no reponse. So, I can set distortion coefficients to [0,0,0,0,0] in your understanding?

hysts commented 3 years ago

@yongwuSHU

Oh, I see. Yes, I think it's OK to assume all distortion coefficients are zero.

yongwuML commented 3 years ago

@hysts Thank you very much.