This article explains the differences between Conv2D and Conv3D networks and provides a TensorFlow 2/Keras based implementation of a Conv3D network using the 3D MNIST dataset. It discusses the architecture and training of the model, as well as evaluating its performance.
Table of contents
Conv2D vs Conv3DToday’s dataset: 3D MNISTCreating the modelModel importsModel configurationAdding helper functionsData PreparationModel architecture and trainingModel EvaluationThe model all togetherSummaryReferencesSort: