PoseCNN: A Convolutional Neural Network for 6D Object Pose Estimation in Cluttered Scenes

1 Nov 2017Yu XiangTanner SchmidtVenkatraman NarayananDieter Fox

Estimating the 6D pose of known objects is important for robots to interact with the real world. The problem is challenging due to the variety of objects as well as the complexity of a scene caused by clutter and occlusions between objects... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
6D Pose Estimation using RGB Occlusion LineMOD Ours PoseCNN+ICP Mean ADD 78 # 2
6D Pose Estimation using RGB YCB-Video PoseCNN Accuracy (ADD) 21.3% # 2
Mean ADD 53.7 # 2
Mean ADD-S 75.9 # 1
6D Pose Estimation using RGBD YCB-Video PoseCNN (ICP) Mean ADD 79.3 # 2
6D Pose Estimation using RGBD YCB-Video ALL PoseCNN+ICP Mean ADD-S 93 # 1
6D Pose Estimation YCB-Video PoseCNN+ICP ADDS AUC 93.0 # 3