Submitted by Sun Jiahao.

Submission data

Full nameSuperpoint Transformer for 3D Scene Instance Segmentation
Publication titleSuperpoint Transformer for 3D Scene Instance Segmentation
Publication authorsSun Jiahao, Qing Chunmei, Tan Junpeng, Xu Xiangmin
Publication venueAAAI 2023
Publication URLhttps://arxiv.org/abs/2211.15766
Input Data TypesUses Color,Uses Geometry        Uses 3D
Programming language(s)Python with CUDA
HardwareRTX 3090
Source code or download URLhttps://github.com/sunjiahao1999/SPFormer
Submission creation date17 Apr, 2022
Last edited30 Nov, 2022

3D semantic instance results



Infoavg ap 50%bathtubbedbookshelfcabinetchaircountercurtaindeskdoorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwindow
permissive0.7700.9030.9030.8060.6090.8860.5680.8150.7050.7110.6550.6520.6851.0000.7890.8090.7761.0000.583