AI Enthusiast Creates The LEGO Sorting Machine Using Neuronet
This December AI enthusiast and game developer Daniel West has claimed to create the world’s first Universal LEGO Sorting Machine that uses AI algorithm:
“Using AI, this machine can recognize and sort any LEGO part that has ever been produced”
Later he has also revealed the AI algorithm basics behind the LEGO Sorting Machine and it mechanics, highlighting that the key part is its capture unit with the camera.
The machine built from LEGO pieces itself uses a neural network to match LEGO pieces with their 3D images. To train the neural network, Daniel had to create over 100 00 labeled LEGO Training Images.
Daniel West has been working on this project for about 2 years. Although this might not be the first project of this kind, its efficiency and elegance are stunning – it takes about 2 seconds to sort one LEGO piece. The creator also aims to submit an academic paper on his research and considers to publish the AI source code to foster other developments.