This project was an application of 1 shot 3d object generation, and was able to take a 2nd photo of an arduino board, collect the importatn data and generate a kicad model. It then is able to take the generated model, and retrain the currently deployed CV model and is abke to fully figure out in 3d space and orientation, and is able to then run a classical ML model on physics and grasp orientation that gets deployed for cuda based motion and planning for manipulation.
Log in or sign up for Devpost to join the conversation.