Submission: CosyPose-ECCV20-SYNT+REAL-8VIEWS/HB

Download submission
Submission name
Submission time (UTC) Aug. 18, 2020, 7:17 p.m.
User yann_labbe
Task Model-based 6D localization of seen objects
Dataset HB
Training model type Default
Training image type Synthetic (only PBR images provided for BOP Challenge 2020 were used)
Description
Evaluation scores
AR:0.746
AR_MSPD:0.804
AR_MSSD:0.744
AR_VSD:0.691
average_time_per_image:0.427

Method: CosyPose-ECCV20-SYNT+REAL-8VIEWS

User yann_labbe
Publication Labbé et al, CosyPose: Consistent multi-view multi-object 6D pose estimation, ECCV 2020
Implementation https://github.com/ylabbe/cosypose
Training image modalities RGB
Test image modalities RGB
Description

Full multi-view method described in the paper, ran with 8 views and using detections and pose estimation models of CosyPose-ECCV20-SYNT+REAL-1VIEW (slightly different from the paper).

Computation time for each group of image is divided by the number of images.

Computer specifications CPU: 20-core Intel Xeon 6164 @ 3.2 GHz, GPU: Nvidia V100