Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Computing | en_US |
| dc.contributor.advisor | Chan, Keith (COMP) | - |
| dc.creator | Wong, Chai Kwok | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/9403 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | - |
| dc.rights | All rights reserved | en_US |
| dc.title | Classifying fashion styles using deep learning | en_US |
| dcterms.abstract | Deep learning, or deep neural network, is a subclass of machine learning that trains layers of interconnected neurons for different applications. In which, convolution neural network (CNN) and its variations are typically used in image applications like classification and object detection. In this dissertation, we evaluated the application of deep learning on fashion analytics, especially fashion attribute annotation, using the trick of re-training final layers of existing state-of-the-art CNN models. We adopted the Inception V3 model and faster RCNN model with RESNET101 classifier to train the computer to classify fashion images and learn to detect objective and subjective fashion features. The evaluated classifier structures are the single labelled single CNN model, the multi labelled binary CNN models, the multi labelled single CNN model, the faster R-CNN model and a combined model of R-CNN with multi labelled CNN model. We evaluated the training efficiencies, sample requirement, training sample grouping, training accuracies and recall of the models. The faster R-CNN structure evaluated was typically applied on object/item detection, so we believe the application of R-CNN on local and global fashion style attribute detection is a novel attempt which achieved better results than conventional whole image approaches by CNNs in our study. We also discussed the application of the above models to build a machine for fashion scoring as well as application of the CNN model in a fashion design process. | en_US |
| dcterms.extent | v, 78 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2018 | en_US |
| dcterms.educationalLevel | M.Sc. | en_US |
| dcterms.educationalLevel | All Master | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.LCSH | Machine learning | en_US |
| dcterms.LCSH | Image processing -- Digital techniques | en_US |
| dcterms.accessRights | restricted access | en_US |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 991022109836903411.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 5.69 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/9403