Penerapan Swin Transformer Pada Klasifikasi Bentuk Wajah Untuk Rekomendasi Bentuk Kacamata

Maulana, Alif (2025) Penerapan Swin Transformer Pada Klasifikasi Bentuk Wajah Untuk Rekomendasi Bentuk Kacamata. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) organized (1)_organized.pdf Download (1MB)
	Text (BAB 1) SKRIPSI PERPUS-22-25.pdf Download (303kB)
	Text (BAB 2) SKRIPSI PERPUS-26-43.pdf Restricted to Repository staff only until 17 September 2028. Download (716kB)
	Text (BAB 3) SKRIPSI PERPUS-44-72.pdf Restricted to Repository staff only until 17 September 2028. Download (1MB)
	Text (BAB 4) SKRIPSI PERPUS-73-116.pdf Restricted to Repository staff only until 17 September 2028. Download (2MB)
	Text (BAB 5) SKRIPSI PERPUS-117-118.pdf Download (295kB)
	Text (Daftar Pustaka) SKRIPSI PERPUS-119-122.pdf Download (241kB)

Abstract

Facial shape serves as an identifying feature that reflects the unique physical characteristics of each individual. In general, facial shapes are categorized into five main types: square, round, oval, heart, and oblong. Information about facial shape can be utilized in various aspects of personalization, including the selection of eyeglass frames. Therefore, the compatibility between eyeglass frame shape and facial shape plays an important role as an aesthetic element that enhances appearance and visual appeal. This research implements the Swin Transformer model to classify facial shapes, where the classification results are used to provide recommendations for eyeglass frame shapes. Swin Transformer is a deep learning architecture that employs a hierarchical image patching approach and processes images through a series of transformer blocks in stages. The results of this study indicate that the model performs well in classification tasks. With a configuration of learning rate 0.0001, batch size 32, and 64 epochs, the model achieved a training accuracy of 99.15%, validation accuracy of 98.47%, and testing accuracy of 97.62%. The small accuracy gap (0.68%) indicates that the model did not experience significant overfitting. Furthermore, the implemented system is capable of performing facial shape classification as well as generating eyeglass frame recommendations. End-to-end testing using 50 test images showed that the classification outputs could be consistently mapped into eyeglass frame recommendations according to the designed rules, with an accuracy of 94%. Thus, this research demonstrates that the developed system is not only accurate in classifying facial shapes but can also be effectively integrated with the eyeglass frame recommendation module.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Wahanani, Henni Endah	NIDN0022097811	henniendah.if@upnjatim.ac.id
Thesis advisor	Al Haromainy, Muhammad Muharrom	NIDN070106503	muhammad.muharrom.if@upnjatim.ac.id

Subjects:

T Technology > T Technology (General)

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Alif Maulana

Date Deposited:

02 Dec 2025 08:23

Last Modified:

02 Dec 2025 09:02

URI:

https://repository.upnjatim.ac.id/id/eprint/43727