Data classification using Convolutional Neural Network
Bukatoi, Aleksandr (2019)
Bukatoi, Aleksandr
2019
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-201904296726
https://urn.fi/URN:NBN:fi:amk-201904296726
Tiivistelmä
The author’s aim in this paper was to understand how deep learning can be connected to automation engineering and which solution would be best suited for data classification.
Based on the research project concluded here, Convolutional Neural Network was chosen as the most suitable solution for the aforementioned task. The goal was to create a network, which would be able to classify images based on features extracted and memorized by the program. The Python programming language and the Anaconda environment were chosen as the most user-friendly environment to demonstrate how the neural network operates.
In order to write this thesis, the author studied a new programming language, examined deep learning and related topics, and was supervised by Raine Lehto throughout all the steps in the project.
At the end of this work the author has achieved the desired results in the form of program classifying images in two categories with a possibility of improving the system for future projects.
Based on the research project concluded here, Convolutional Neural Network was chosen as the most suitable solution for the aforementioned task. The goal was to create a network, which would be able to classify images based on features extracted and memorized by the program. The Python programming language and the Anaconda environment were chosen as the most user-friendly environment to demonstrate how the neural network operates.
In order to write this thesis, the author studied a new programming language, examined deep learning and related topics, and was supervised by Raine Lehto throughout all the steps in the project.
At the end of this work the author has achieved the desired results in the form of program classifying images in two categories with a possibility of improving the system for future projects.