search

ShuhongLL / SGS-SLAM

[ECCV 2024] SGS-SLAM: Semantic Gaussian Splatting For Neural Dense SLAM
BSD 3-Clause "New" or "Revised" License
38 stars 6 forks source link

Question about post_slam_opt.py #1

Open Yingsheng123 opened 2 hours ago

Yingsheng123 commented 2 hours ago

Why do need to execute two files，post_slam_opt.py and slam.py？I find that there is optimization process in slam.py, what is the difference with the other file?

ShuhongLL commented 2 hours ago

This has been adapted from SplaTAM, a post-optimization process after running the SLAM system to enhance map reconstruction quality. Similar modules have also been implemented in previous studies like MonoGS or Gaussian-SLAM, where are defined as a color refinement process. While optional, it is recommended to carry out this step :)