Neural networks are a tool for machine learning in which a computer tries to recognize relationships in a set of data through a process that mimics the way the human brain works. Machine learning is a scientific discipline that is based on designing an algorithm that will use a set of data to learn on specific examples (Naqa and Murphy, 2015). Convolutional neural networks are a subset of neural networks most commonly used for training image-related models. This paper explores the idea of ​​geometric body classification using EEG data. EEG data is shown in voltage flow, and for a better representation of the brain it is better to use more EEG signal measurements, i.e. more electrodes that are placed all around the surface of the brain. Recording of EEG data was done using the OpenBCI Cyton + Daisy device, where the data was recorded with one subject. In order for the data to be as clean and relevant as possible, the room has weaker lighting and there is no penetration of noise or other external disturbances. After recording the dataset, it was processed to be compatible with the Conv2D layer. The source code is available on GitHub: https://github.com/David-Sajina/OpenBCI