Submitted by Liyii Luo.

Submission data

Full nameWeakly Supervised 3D Scene Parsing with Viewpoint Bottleneck
Input Data TypesUses Color,Uses Geometry        Uses 3D
Programming language(s)python with CUDA
HardwareGeForce RTX 3090
Submission creation date18 Jun, 2021
Last edited19 Jun, 2021

3D semantic label results with limited reconstructions



Infoavg ioubathtubbedbookshelfcabinetchaircountercurtaindeskdoorfloorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwallwindow
0.2560.0000.4790.3770.2040.5510.2050.2190.2350.2240.9030.0920.0880.1220.0000.0030.3540.3540.0000.6760.034