Preparing a dataset for training and testing neural networks in MATLAB

Abstract

Preparing a dataset for training and testing neural networks in MATLAB

Kovalev A.V., Isaeva A.S.

Incoming article date: 22.12.2021

The paper proposes a method for preparing a dataset for training and testing neural networks in the MATLAB environment using the Image Processing Toolbox package of functions. An example of the implementation of this technique for preparing a dataset for training convolutional neural networks to classify wheat seeds into high-quality and low-quality classes according to its images is given. This technique was used to obtain a dataset of 900 images, which were successfully used to train and test the convolutional neural network.

Keywords: dataset, convolutional neural networks, artificial intelligence, image classification