vggt icon indicating copy to clipboard operation
vggt copied to clipboard

about two-view matching

Open miaolin968 opened this issue 6 months ago • 1 comments

Dear author: In Table 4 of your paper, you presented the results of Two-View matching. Specifically, how was the Two-View matching implemented? Were the query_points modified to include all pixel poses?

miaolin968 avatar Jun 07 '25 15:06 miaolin968

Hi,

We just fed two frames into the network, together with query points. The query points are the keypoints from ALIKED.

jytime avatar Jun 10 '25 13:06 jytime