Submitted by Angela Dai.

Submission data

Full nameScanNet semantic segmentation network
Publication titleScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes
Publication authorsAngela Dai, Angel X. Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, Matthias Nie├čner
Publication venueCVPR'17
Publication URLhttp://www.scan-net.org/
Input Data TypesUses Geometry        Uses 3D
Programming language(s)pytorch
HardwareGTX 1080
Websitehttp://www.scan-net.org/
Source code or download URLhttps://github.com/scannet/scannet
Submission creation date4 Jul, 2018
Last edited14 Jul, 2018

3D semantic label results

Infoavg ioubathtubbedbookshelfcabinetchaircountercurtaindeskdoorfloorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwallwindow
permissive0.3060.2030.3660.5010.3110.5240.2110.0020.3420.1890.7860.1450.1020.2450.1520.3180.3480.3000.4600.4370.182