search

xy-guo / GwcNet

Group-wise Correlation Stereo Network, CVPR 2019
MIT License
322 stars 62 forks source link

Test image size problem #21

Open QingWindIsStillTheWind opened 4 years ago

QingWindIsStillTheWind commented 4 years ago

I would like to ask the author, if the size of the image used for network training is 960x540, then when I want to use this network test, the resolution of the photos taken by my camera is very large, such as 4608x3456. When I take this picture to test, the parallax map effect is very bad. When I reduce the original image to 960x540, the parallax map effect is very good OK, but I don't know how to restore the parallax value of the original image. Or how to use your network to test higher resolution images without enough similar data sets to train? Hope that the author can guide, thank you very much!

xy-guo commented 3 years ago

You can compute the disparity map using 960x540. Once you got the output disparity map, resize it back to 4608x3456. Then multiply a scale factor 4.8 (4608 / 960)