search

ubc-vision / COTR

Code release for "COTR: Correspondence Transformer for Matching Across Images"(ICCV 2021)
Apache License 2.0
460 stars 58 forks source link

How to get depth map #27

Closed shizukasonn closed 2 years ago

shizukasonn commented 2 years ago

Hi, thank you for your nice work:) Sorry I'm new to this filed, I wonder how can I get the depth map of two images using COTR results? thanks a lot

jiangwei221 commented 2 years ago

Hi, to obtain the depth map, you need to triangulate the 3D points using the correspondences, which requires known camera poses and camera intrinsics. Once the 3D points are obtained, you can project those points to the two cameras and obtain 2 depth maps. There is a demo program in the repo that shows how to do this 3D points triangulation.