الفهرس 
ContentsOnly 14 pages are availabe for public view
 |  | from | 138 |  |  |
| | | from | 138 | | |
Abstract
The development of automated classification schemes for galaxy images is important to identify, classify, and study the formation and subsequent evolution of galaxies and their content in our universe and this is one of the most important challenges and intensive research area facing astronomers today. By classifying galaxies into different categories, scientists can build a deeper understanding of their form and evolve. This thesis proposed two robust automated intelligent systems based on; (i) Machine Learning (ML) techniques, and (ii) Fusion between Neutrosophic techniques (NTs) and ML techniques for galaxy morphology astronomical classification to classify different types of galaxies images (Hubble types) based on its features into three general categories: (i) Elliptical galaxies, (ii) Spiral galaxies, and (iii) Irregular galaxies. The performance of the NTs and ML based classifiers was evaluated, based on a selected set of features. They are a combination of a set of Morphic Features (MFs); derived from image analysis and Principal Component Analysis (PCA) features. These features are combined and arranged to constitute five groups of features. Then, the NTs were applied on these features to get three robust components. The results showed that; the combination between the NTs and Multilayer Perceptron (MLP) based classifier for MFs with 4PCs gives the best results; Mean-Square Error )MSE( = 0.0001; Normalized Mean Square Error )NMSE( = 0.0009; Correlation Coefficient )r( = 0.9997, and Error = 0.4212 with an accuracy about 99.5788 % in total among all tested ML classifiers for all groups of selected features and the accuracy of each type is given as follows: elliptical, 99.7899%; spiral, 99.8989%; irregular, 99.89% for a three robust components of NTs for a features sets of nine Morphic Features, and 4 Principal Components features. The testing accuracy was compared with those of other related works and the results showed the high performance of the proposed method. It is also concluded that a small set of features is sufficient to classify galaxy images and provide a fast classification. In addition, our model algorithm can be applied to large-scale galaxy classification in forthcoming surveys. The experimental results are performed based on a sample from EFIGI (Extraction de Formes Id´ealis´ees de Galaxies en Imagerie) catalog. The results were verified using performance metrices. The proposed architecture was implemented using a software package (MATLAB R 2017b). The implementation was CPU Specific and run-in windows environment with 64bit support. All experiments were conducted on a server with core i5 Intel processor (2.60 GHz) and 4.00 GB Ram.