Submission: FFB6D/LM/FFB6D-lm-synthetic

Submission name FFB6D-lm-synthetic
Submission time (UTC) Dec. 3, 2021, 3:45 p.m.
User fegorsch
Task Pose estimation (BOP 2019-2022)
Dataset LM
Training model type Default
Training image type Synthetic (provided)
Description Trained for 316600 iterations at 0.5 seconds per iteration, about 7 days.
Evaluation scores
AR:0.643
AR_MSPD:0.709
AR_MSSD:0.702
AR_VSD:0.519
average_time_per_image:0.196

Method: FFB6D

User fegorsch
Publication Yisheng He et al.: FFB6D: A Full Flow Bidirectional Fusion Network for 6D Pose Estimation, IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021
Implementation https://github.com/ethnhe/FFB6D
Training image modalities RGB-D
Test image modalities RGB-D
Description

FFB6D trained with purely synthetic BlenderProc-images.

Computer specifications AMD Ryzen 5 3600X 6-Core Processor, NVIDIA TITAN RTX