Submission: GenFlow-coarse/HB

Download submission
Submission name
Submission time (UTC) Sept. 26, 2023, 7:54 a.m.
User sp9103
Task 6D localization of unseen objects
Dataset HB
Description
Evaluation scores
AR:0.283
AR_MSPD:0.450
AR_MSSD:0.254
AR_VSD:0.144
average_time_per_image:5.108

Method: GenFlow-coarse

User sp9103
Publication Not yet
Implementation -
Training image modalities RGB
Test image modalities RGB
Description

Submitted to: BOP Challenge 2023

Training data: MegaPose-GSO and MegaPose-ShapeNetCore

Onboarding data: No

Used 3D models: Default, CAD

Notes:

In this submission, CNOS_fastSAM [A] detections are used as the input to our pose estimation method. Our pose estimation method uses the coarse-to-fine strategy following the MegaPose [B] structure. A single model is used for all datasets.

Our coarse network is based on the MegaPose [B] coarse network. The main differences from the original MegaPose paper are as follows:

  • To bypass the computation of rendering tons of images, we render fewer images and run the coarse network. Then we create a GMM with the top-k hypotheses and run the coarse network to the sampled hypotheses from the GMM.

[A] Nguyen et al.: CNOS: A Strong Baseline for CAD-based Novel Object Segmentation, arXiv 2023
[B] Labbé et al.: MegaPose: 6D Pose Estimation of Novel Objects via Render & Compare, CoRL 2022

List of contributors: Sungphill Moon (sungphill.moon@naverlabs.com), Hyeontae Son (son.ht@naverlabs.com)

If you have any questions, feel free to contact us.

Computer specifications GPU V100; CPU Intel Xeon Gold6248@2.5G