Vision-Language Models Untuk Rekomendasi Kesesuaian Pakaian Pada Platform Virtual Try-On Kiosk

Kusyuniardi, M Raidan Arsyal Yudeindra (2026) Vision-Language Models Untuk Rekomendasi Kesesuaian Pakaian Pada Platform Virtual Try-On Kiosk. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5024211053-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Platform Virtual Try-On (VTO) konvensional umumnya bersifat reaktif, mengharuskan pengguna untuk menelusuri katalog secara manual. Keterbatasan ini mengurangi potensi personalisasi dan efisiensi dalam pengalaman berbelanja. Penelitian ini bertujuan untuk mengembangkan sebuah sistem VTO Kiosk yang cerdas dan proaktif dengan memanfaatkan VisionLanguage Models(VLM). Fokus utama dari penelitian ini adalah merancang sistem yang mampu melakukan Visual Style Matching, yaitu menganalisis gaya berpakaian pengguna secara realtime melalui input kamera, dan Prediksi Kesesuaian Pakaian, yaitu merekomendasikan produk dari katalog yang cocok dengan gaya yang telah teridentifikasi. Metodologi yang digunakan adalah pengembangan eksperimental, mencakup perancangan arsitektur sistem kios, pemanfaatan pre-trained VLM berbasis instruksi pada dataset fashion, dan pengujian fungsionalitas secara end-to-end. Hasil yang diharapkan adalah sebuah prototipe VTO Kiosk yang mampu memberikan rekomendasi fashion yang dipersonalisasi secara otomatis berdasarkan penampilan visual pengguna, sehingga menciptakan interaksi belanja yang lebih cerdas, intuitif, dan menarik.
==================================================================================================================================
Conventional Virtual Try-On (VTO) platforms are generally reactive, requiring users to manually browse catalogs. This limitation reduces the potential for personalization and efficiency in the shopping experience. This research aims to develop an intelligent and proactive VTO Kiosk system by leveraging Vision-Language Models (VLM). The main focus of this research is to design a system capable of performing Visual Style Matching, which involves
analyzing a user’s clothing style in real-time via camera input, and Clothing Compatibility Prediction, which recommends products from a catalog that match the identified style. The methodology employed is experimental development, encompassing the design of the kiosk system architecture, utilization of pre-trained instruction-based VLM on a fashion dataset, and end-to-end functional testing. The expected outcome is a VTO Kiosk prototype capable of automatically providing personalized fashion recommendations based on the user’s visual appearance, thereby creating a more intelligent, intuitive, and engaging shopping interaction

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Vision-Language Model(VLM),Prompt Engineering, Virtual Try-On, Rekomendasi Pakaian, Evaluasi Heuristik, Vision-Language Model (VLM), Prompt Engineering, Virtual Try-On, Clothing Recommendation, Heuristic Evaluation.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.585 Cloud computing. Mobile computing. Q Science > QA Mathematics > QA76.758 Software engineering Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9 Computer algorithms. Virtual Reality. Computer simulation. Q Science > QA Mathematics > QA76.9.U83 Graphical user interfaces. User interfaces (Computer systems)--Design. T Technology > T Technology (General) > T385 Visualization--Technique T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems
Divisions:	Faculty of Electrical Technology > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User:	M. Raidan Arsyal Yudeindra Kusyuniardi
Date Deposited:	31 Jul 2026 05:23
Last Modified:	31 Jul 2026 05:23
URI:	http://repository.its.ac.id/id/eprint/140452

Actions (login required)

View Item