In this repository there will be code modified from the Tensorflow CNN example (CIFAR10) at https://www.tensorflow.org/tutorials/deep_cnn.
Objective: to test the concept of a 3D search engine, a machine learning program which can divide parts of a large aseembly and recognize individual pieces.
To test this concept we wil train our networks on images of 10 different LEGO parts, and hopefully classify larger assemblies made from those 10 parts.