Juan Ricardo Wilches Cortina is the primary author and contributor to this project.
In this project, we reused the code provided by the authors of Deep Object Pose Estimation for Semantic Robotic Grasping of Household Objects.
The original code can be found in the following Github.
-
First, we performed a literature review in which we found that the state-of-the-art approaches are deep learning based.
-
Second, we replicated the results of the DOPE paper with the weights published by the authors.
-
Third, we downloaded the YCB-Video benchmark dataset published in BOP: Benchmark for 6D Object Pose Estimation. Finally, we evaluated the performance of the DOPE model ourselves using the ADD metric proposed in PoseCNN: A Convolutional Neural Network for 6D Object Pose Estimation in Cluttered Scenes. We had to use the original object models in order to compare the ground truth versus the estimation.