Closed ghoshaw closed 2 years ago
We haven't tried that. Did you use the pretrained model? Maybe the image resolution is different? The pretrained Coco model performed quite well on KITTI ...
Yes, I used the pretrained model. And I found I used some wrong parameter so the result is bad. It is solved anyway. Thanks.
Hi, I used kp2d as keypoint and local descriptor and netvlad as global descriptor in hloc for Aachen dataset. The result is very bad. Has anyone done visual location using the method ?
Hello,I use the dockerfile to build the environment,but the docker file doesn't work well;it shows me that
Could you please share your docker file with me? @ghoshaw @RaresAmbrus thank you very much
Hi, I used kp2d as keypoint and local descriptor and netvlad as global descriptor in hloc for Aachen dataset. The result is very bad. Has anyone done visual location using the method ?