ACADSTAFF UGM

CREATION
Title : Classification of Galaxy Morphological Image Based on Convolutional Neural Network
Author :

Wahyono, Ph.D. (1) MUHAMMAD ARIF RAHMAN (2) Dr. Azhari, MT. (3)
Date : 0 2018
Keyword : image processing, convolutional neural network, elliptical galaxy, spiral galaxy, irregular galaxy image processing, convolutional neural network, elliptical galaxy, spiral galaxy, irregular galaxy
Abstract : Astronomers use the term ‘morphology’ to refer to the structural properties of galaxies. The formation of galaxy involves a complex combination of effects, namely, radioactive cooling, star formation, merging of foreign celestial bodies, and etc. By classifying galaxies into different categories, scientists can build a deeper understanding of they form and evolve and even make an estimation of the amount of time that had passed since the ‘Big Bang’ until the present. This research aims to classify images of different types of galaxies into three more general categories: Elliptical, Spiral and Irregular. The research is based on convolutional neural network, specially using inception. The total number of images that were used in this research were 206 images of elliptical galaxies, 320 images of spiral galaxies and 184 images of irregular galaxies. The test result showed that images that went through image processing showed a rather poor testing accuracy compared to not using any form of image processing. The best testing accuracy that this research obtained was 78.3%.
Group of Knowledge : Ilmu Komputer
Original Language : English
Level : Internasional
Status :
Published
Document
No Title Document Type Action
1 Full Dok PAK Classification of Galaxy Morphological Image Based on.pdf
Document Type : [PAK] Full Dokumen
[PAK] Full Dokumen View
2 05--Check Simliarity Classification of Galaxy Morphological Image Based on Convolutional Neural Network.pdf
Document Type : [PAK] Cek Similarity
[PAK] Cek Similarity View
3 jurnal_2046325_63ca7ac1028089174aa5bfcdad5ce0e1.pdf
Document Type : [PAK] Full Dokumen
[PAK] Full Dokumen View