Method: GigaPose+GenFlow+kabsch (5 hypothesis)

User sp9103
Publication Genflow: Generalizable recurrent flow for 6d pose refinement of novel objects, CVPR 2024
Implementation -
Training image modalities RGB-D
Test image modalities RGB-D
Description

We use GigaPose [A] to extract 5 hypothesis and run GenFlow[B].

[A] Nguyen et al.: GigaPose: Fast and Robust Novel Object Pose Estimation via One Correspondence, CVPR 2024.

[B] Genflow: Generalizable recurrent flow for 6d pose refinement of novel objects, CVPR 2024

Computer specifications V100

Public submissions

Date Submission name Dataset
2023-11-30 06:44 - LM-O
2023-11-30 06:45 - T-LESS
2023-11-30 06:45 - TUD-L
2023-11-30 06:46 - IC-BIN
2023-11-30 06:46 - HB
2023-11-30 06:47 - ITODD
2023-11-30 06:55 - YCB-V
2024-09-18 10:09 - LM-O
2024-09-18 10:10 - T-LESS
2024-09-18 10:10 - TUD-L
2024-09-18 10:10 - IC-BIN
2024-09-18 10:11 - ITODD
2024-09-18 10:11 - HB
2024-09-18 10:11 - YCB-V