*1.3. Challenge*

Although great progress has been made in the research of 6D pose estimation in recent years, there are still some challenging problems to be solved in practical application. When the background is messy, the viewpoint and illumination change greatly, or the scene has less texture, the accuracy and the robustness of 6D pose estimation needs to be improved. Learning-based approaches are relatively robust in these conditions. However, 2D-information-based approaches and 3D-information-based approaches do not perform well. However, for learning-based approaches, adequate training data and training time are required, and challenges related to the requirement of offline training and practicality are also problems to be solved.
