Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
A hybrid deep learning approach for texture analysis
Texture classification is a problem that has various applications such as remote sensing and forest species recognition. Solutions tend to be custom fit to the dataset used but fails to generalize. The Convolutional Neural Network (CNN) in combination with Support Vector Machine (SVM) form a robust...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Published: |
AUC Knowledge Fountain
2016
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613409493647361 |
|---|---|
| access_status_str | Open Access |
| author | Adly, Hussein |
| author_browse | Adly, Hussein |
| author_facet | Adly, Hussein |
| author_sort | Adly, Hussein |
| collection | Thesis |
| dc_rights_str_mv | The author retains all rights with regard to copyright. The author certifies that written permission from the owner(s) of third-party copyrighted matter included in the thesis, dissertation, paper, or record of study has been obtained. The author further certifies that IRB approval has been obtained for this thesis, or that IRB approval is not necessary for this thesis. Insofar as this thesis, dissertation, paper, or record of study is an educational record as defined in the Family Educational Rights and Privacy Act (FERPA) (20 USC 1232g), the author has granted consent to disclosure of it to anyone who requests a copy. |
| description | Texture classification is a problem that has various applications such as remote sensing and forest species recognition. Solutions tend to be custom fit to the dataset used but fails to generalize. The Convolutional Neural Network (CNN) in combination with Support Vector Machine (SVM) form a robust selection between powerful invariant feature extractor and accurate classifier. The fusion of classifiers shows the stability of classification among different datasets and slight improvement compared to state of the art methods. The classifiers are fused using confusion matrix after independent training of each using the same training set, then put to test. Statistical information about each classifier is fed to a confusion matrix that generates two confidence measures used in building two binary classifiers. The binary classifier is allowed to activate or deactivate a classifier during testing time based on a confidence measure obtained from the confusion matrix. The method obtained results approaching state of the art with a difference less than 1% in classification success rates. Moreover, the method was able to maintain this success rate among different datasets while other methods had failed to obtain similar stability. Two datasets had been used in this research Brodatz and Kylberg where the results came 98.17% and 99.70%. In comparison to conventional methods in the literature, it came as 98.9% and 99.64% respectively. |
| format | Thesis |
| id | oai:fount.aucegypt.edu:etds-1358 |
| institution | American University in Cairo (Egypt) |
| last_indexed | 2026-06-10T12:35:41.195Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from AUC Knowledge Fountain — bepress |
| publishDate | 2016 |
| publishDateRange | 2016 |
| publishDateSort | 2016 |
| publisher | AUC Knowledge Fountain |
| publisherStr | AUC Knowledge Fountain |
| record_format | dspace |
| source_str | AUC Knowledge Fountain — bepress |
| spelling | oai:fount.aucegypt.edu:etds-1358 A hybrid deep learning approach for texture analysis Adly, Hussein Texture classification is a problem that has various applications such as remote sensing and forest species recognition. Solutions tend to be custom fit to the dataset used but fails to generalize. The Convolutional Neural Network (CNN) in combination with Support Vector Machine (SVM) form a robust selection between powerful invariant feature extractor and accurate classifier. The fusion of classifiers shows the stability of classification among different datasets and slight improvement compared to state of the art methods. The classifiers are fused using confusion matrix after independent training of each using the same training set, then put to test. Statistical information about each classifier is fed to a confusion matrix that generates two confidence measures used in building two binary classifiers. The binary classifier is allowed to activate or deactivate a classifier during testing time based on a confidence measure obtained from the confusion matrix. The method obtained results approaching state of the art with a difference less than 1% in classification success rates. Moreover, the method was able to maintain this success rate among different datasets while other methods had failed to obtain similar stability. Two datasets had been used in this research Brodatz and Kylberg where the results came 98.17% and 99.70%. In comparison to conventional methods in the literature, it came as 98.9% and 99.64% respectively. 2016-02-01T08:00:00Z thesis application/pdf https://fount.aucegypt.edu/etds/359 https://fount.aucegypt.edu/context/etds/article/1358/viewcontent/Thesis.pdf The author retains all rights with regard to copyright. The author certifies that written permission from the owner(s) of third-party copyrighted matter included in the thesis, dissertation, paper, or record of study has been obtained. The author further certifies that IRB approval has been obtained for this thesis, or that IRB approval is not necessary for this thesis. Insofar as this thesis, dissertation, paper, or record of study is an educational record as defined in the Family Educational Rights and Privacy Act (FERPA) (20 USC 1232g), the author has granted consent to disclosure of it to anyone who requests a copy. Theses and Dissertations AUC Knowledge Fountain classification NA NA |
| spellingShingle | classification NA NA Adly, Hussein A hybrid deep learning approach for texture analysis |
| title | A hybrid deep learning approach for texture analysis |
| title_full | A hybrid deep learning approach for texture analysis |
| title_fullStr | A hybrid deep learning approach for texture analysis |
| title_full_unstemmed | A hybrid deep learning approach for texture analysis |
| title_short | A hybrid deep learning approach for texture analysis |
| title_sort | hybrid deep learning approach for texture analysis |
| topic | classification NA NA |
| url | https://fount.aucegypt.edu/etds/359 https://fount.aucegypt.edu/context/etds/article/1358/viewcontent/Thesis.pdf |
| work_keys_str_mv | AT adlyhussein ahybriddeeplearningapproachfortextureanalysis AT adlyhussein hybriddeeplearningapproachfortextureanalysis |