Submitted by zhengzhe liu.

Submission data

Full nameOne_Thing_One_Click
DescriptionOnly use 0.02% annotation.
Publication titleOne Thing One Click: A Self-Training Approach for Weakly Supervised 3D Semantic Segmentation
Publication authorsZhengzhe Liu, Xiaojuan Qi, Chi-Wing Fu
Publication venueCVPR 2021
Publication URLhttps://arxiv.org/abs/2104.02246
Input Data TypesUses Color,Uses Geometry        Uses 3D
Programming language(s)python cuda
Hardware2080 ti
Websitehttps://github.com/liuzhengzhe/One-Thing-One-Click
Source code or download URLhttps://github.com/liuzhengzhe/One-Thing-One-Click
Submission creation date3 Jun, 2022
Last edited3 Jun, 2022

3D semantic instance results



Infoavg ap 50%bathtubbedbookshelfcabinetchaircountercurtaindeskdoorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwindow
permissive0.5290.6670.7180.7770.3990.6830.0000.6690.1380.3910.3740.5390.3600.6410.5560.7740.5930.9970.251