Heng Zhou, Hongyu Wang, Shuhong Liu, Mingrui Li, Na Cheng, Tianchen Deng, Guohao Zhu
We lifted 2 functions out of this paper's own repositories and ran 1 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| shuhongll/sgs-slam | — | 1 of 2 |
| Function | Status | Where it lives |
|---|---|---|
| build_rotation | Ran | shuhongll/sgs-slam/utils/gs_helpers.py code served (permissive licence) · get_code("20a77144ffa65865") |
| transform_to_frame | Not yet run | shuhongll/sgs-slam/utils/gs_helpers.py code served (permissive licence) · get_code("c923f56d82bd5eff") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
We present SGS-SLAM, the first semantic visual SLAM system based on Gaussian Splatting. It incorporates appearance, geometry, and semantic features through multi-channel optimization, addressing the oversmoothing limitations of neural implicit SLAM systems in highquality rendering, scene understanding, and object-level geometry. We introduce a unique semantic feature loss that effectively compensates for the shortcomings of traditional depth and color losses in object optimization. Through a semantic-guided keyframe selection strategy, we prevent erroneous reconstructions caused by cumulative errors. Extensive experiments demonstrate that SGS-SLAM delivers state-of-the-art performance in camera pose estimation, map reconstruction, precise semantic segmentation, and object-level geometric accuracy, while ensuring real-time rendering capabilities.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2402.03246")
get_code_for_paper("2402.03246")
have("2402.03246")
Connect an agent — have() is free.