LEGO-Net is a data-driven transformer-based iterative method for LEarning reGular rearrangement of Objects in messy rooms. LEGO-Net is partly inspired by
diffusion models – it starts with an initial messy state and iteratively “de-noises” the position and orientation of objects to a regular state while reducing distance traveled.
In this work, leveraging the latest development in machine learning and robot learning, we develop an image-based sorting system for tackling this challenging task.
The results indicate that some occurrences of undermelting, overmelting, and material spatter can be detected that can then be correlated to localized defects, delamination, and layer separation.