Method: KeyMatchNet

User frederikhagelskjaer
Publication KeyMatchNet: Zero-Shot Pose Estimation in 3D Point Clouds by Generalized Keypoint Matching
Training image modalities D
Test image modalities D

Submitted to: BOP Challenge 2023

Training data: Trained on the synthetic point clouds of electrical components in bins.

Onboarding data: Only for Task 4.

Used 3D models: Default.

Notes: Altered to run with multiple pose estimations per image. Use the default detections as the network has not been trained with a detector.

Computer specifications Two NVIDIA GeForce RTX 2080 GPUs. With an Intel(R) Core(TM) i9-9820X CPU @ 3.30GHz CPU.

Public submissions

No submissions yet.