Virtual Personalized Fashion Styling Assistant for Online Platforms

Sayuri Jayarathne

Keywords: Augmented reality, Machine learning, Virtual try on, Fashion insight navigator, Conversational AI, Fabric advisor

Abstract: Online shopping for fashion has a high rate of returns due to size mistakes and the inability to try on the clothes before buying. This research presents the Virtual Personalized Fashion Styling Assistant (VPFSA), an artificial intelligence (AI) system integrating augmented reality (AR), machine learning (ML), and natural language processing (NLP) to enhance online shopping. The VPFSA consists of four main components: (1) a Fashion Insight Navigator that forecasts trends through the application of a Random Forest Classifier, (2) a Smart Fabric Advisor that comes up with fabric insights via the T5-Large model, (3) a Virtual Try-On system that applies OpenCV, Lens Studio, and TensorFlow to facilitate real-time 3D visualization of apparel, and (4) a Virtual Styling Assistant that fine-tunes the random forest classifier model to make personalized suggestions. Developed with React, Python, and cloud hosting, VPFSA improves accuracy of fit, prevents returns, and boosts customer interaction. Dynamically adapting to the user's body measurements and preferences, it offers an e-commerce solution that is scalable, transforming online fashion shopping while being sustainable.

Submission Number: 1