You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello,
First, thanks for the efforts for the datasets.
I am trying to use the ground truth optical flow vector to warp the previous and current images. Is there in example code to achieve the same? I have tried the Python code linked here:
Using this code, the output interpolated image looks unsatisfactory, either it has lot of holes using img1 option, as explained in the documentation or using img2, the output image has duplicated overlays at the bottom of the image.
Therefore, it will be nice if there is an example code demonstrating the warped image; and also a documentation explaining the flow field clearly whether
the flow moves each pixel of source(previous frame) image to a location in the destination(current frame). in this case, the output wont be a regular grid and rounding the indices will lead to holes
it specifies for each output pixel(in destination image), where the source pixel comes from in the previous image
I hope my query is clear and if not, i am happy to clarify.
rgds
krishnan
The text was updated successfully, but these errors were encountered:
Hello,
First, thanks for the efforts for the datasets.
I am trying to use the ground truth optical flow vector to warp the previous and current images. Is there in example code to achieve the same? I have tried the Python code linked here:
Using this code, the output interpolated image looks unsatisfactory, either it has lot of holes using img1 option, as explained in the documentation or using img2, the output image has duplicated overlays at the bottom of the image.
Therefore, it will be nice if there is an example code demonstrating the warped image; and also a documentation explaining the flow field clearly whether
I hope my query is clear and if not, i am happy to clarify.
rgds
krishnan
The text was updated successfully, but these errors were encountered: