MT2DRNET: MODEL PENGENALAN CITRA MOTIF SONGKET PALEMBANG BERBASIS TRANSFER LEARNING MENGGUNAKAN REGULARIZATION DAN MODIFIED TOP-HAT TRANSFORM

NOPRISSON, HANDRIE and Ermatita, Ermatita and Abdiansah, Abdiansah (2023) MT2DRNET: MODEL PENGENALAN CITRA MOTIF SONGKET PALEMBANG BERBASIS TRANSFER LEARNING MENGGUNAKAN REGULARIZATION DAN MODIFIED TOP-HAT TRANSFORM. Doctoral thesis, Sriwijaya University.

	Text RAMA_21001_03013682126008.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (16MB) | Request a copy
	Text RAMA_21001_03013682126008_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (19MB) | Request a copy
	Text RAMA_21001_03013682126008_0013096707_0001108401_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (1MB)
	Text RAMA_21001_03013682126008_0013096707_0001108401_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_21001_03013682126008_0013096707_0001108401_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_21001_03013682126008_0013096707_0001108401_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy
	Text RAMA_21001_03013682126008_0013096707_0001108401_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_21001_03013682126008_0013096707_0001108401_06.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (343kB) | Request a copy
	Text RAMA_21001_03013682126008_0013096707_0001108401_07_ref.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (387kB) | Request a copy
	Text RAMA_21001_03013682126008_0013096707_0001108401_08_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (13MB) | Request a copy

Abstract

The current millennial generation is rarely able to recognize the names of the motifs on Palembang's songket woven fabrics. If you want to know a particular motif, the millennial generation usually asks cultural experts or woven fabric centers who understand the history of Palembang's songket woven fabric. Utilization of information technology, especially artificial intelligence technology, can be an option to promote knowledge of motifs on Palembang's songket woven fabrics. One of the artificial intelligence technologies to support this program is pattern recognition. The main objective of this study is to propose a new pattern recognition model, namely MT2DRNET for the introduction of Palembang songket woven fabrics from the fusion results between the DRNET and MT2 models. The issues that will be examined are reducing overfitting in the DRNET model by using the dropout (DR) regularization technique in the ResNet (RNET) model, optimizing the modified top-hat transform (MT2) method from applying the equalized cumulative histogram (ECH) function in the top-hat method. hat transform to improve the image quality of Palembang songket woven fabric, produce a new MT2DRNET model for the introduction of Palembang songket woven fabric from the fusion between the DRNET and MT2 models and produce an application architecture model for the introduction of Palembang songket woven fabric using unified modeling language (UML) and metadata analysis (MDA). The types of woven fabric motifs tested are bintang melati, bunga bintang, bunga mawar, kucing tidur, naga besaung, pucuk rebung balai anak, pucuk rebung penuh and tampuk manggis. Experiments using MT2DRNET have the best performance in terms of training accuracy, validation accuracy, and testing accuracy compared to RNET, DRNET and T2DRNET. MT2DRNET has a training accuracy of 94.86%, validation accuracy of 82.70%, and testing accuracy of 83.50%. The application architecture for the introduction of Palembang songket woven fabric consists of user elements, user interface elements, application module elements (content management and motif recognition) and data management elements. In the recognition module motif, the MT2DRNET model is applied.

Item Type:	Thesis (Doctoral)
Uncontrolled Keywords:	traditional woven fabric, transfer learning, dropout, top-hat transform, equalized cumulative histogram
Subjects:	T Technology > T Technology (General) > T58.5-58.64 Information technology > T58.5 General works Management information systems Cf. HD30.213 Industrial management Cf. HF5549.5.C6+ Communication in personnel management Cf. TS158.6 Automatic data collection systems (Production control)
Divisions:	03-Faculty of Engineering > 21001-Engineering Science (S3)
Depositing User:	Handrie Noprisson
Date Deposited:	28 Nov 2023 06:36
Last Modified:	28 Nov 2023 06:36
URI:	http://repository.unsri.ac.id/id/eprint/131312

Actions (login required)

View Item