The Effect of Segmentation and CNN Architecture in Determining Accuracy Convolutional Neural Network

Manutur Siregar, Herman Mawengkang, Suyanto

2023

Abstract

Convolutional Neural Network is a type of deep learning that is used for image detection or image classification. The images used can be obtained from data banks such as http://www.kaggle.com, the image usually has a different image format, different width and height sizes also in some images contain noise. Segmentation is used to separate the image that becomes information from the noise contained therein. The types of segmentation used are active contour and K- Means and compared to not using segmentation at all. The Convolutional Neural Network architecture was also changed to obtain a better level of accuracy, and to take advantage of the existing CNN architectures such as Alexnet and GoogleNet. From the research conducted, the best accuracy results were obtained from the RGB Image model combined with the GoogleNet model, namely 98.37 K-Means segmentation has better test results when compared to the active contour for classifying lung disease.

Paper Citation

in Harvard Style

Siregar M., Mawengkang H. and Suyanto. (2023). The Effect of Segmentation and CNN Architecture in Determining Accuracy Convolutional Neural Network. In Proceedings of the 3rd International Conference on Advanced Information Scientific Development - Volume 1: ICAISD; ISBN 978-989-758-678-1, SciTePress, pages 46-51. DOI: 10.5220/0012441500003848

in Bibtex Style

@conference{icaisd23,
author={Manutur Siregar and Herman Mawengkang and Suyanto},
title={The Effect of Segmentation and CNN Architecture in Determining Accuracy Convolutional Neural Network},
booktitle={Proceedings of the 3rd International Conference on Advanced Information Scientific Development - Volume 1: ICAISD},
year={2023},
pages={46-51},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012441500003848},
isbn={978-989-758-678-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 3rd International Conference on Advanced Information Scientific Development - Volume 1: ICAISD
TI - The Effect of Segmentation and CNN Architecture in Determining Accuracy Convolutional Neural Network
SN - 978-989-758-678-1
AU - Siregar M.
AU - Mawengkang H.
AU - Suyanto.
PY - 2023
SP - 46
EP - 51
DO - 10.5220/0012441500003848
PB - SciTePress