Abstract:
Fashion recommendation systems do not provide profound personalization and contextual consideration. Their approach often uses static datasets with little or no accommodation for preferred styles, body measurements, or scenario requirements, like weather or occasion, which leads to poor engagement and disillusioned fashion discovery. To address these issues, an AI mobile application was created, Fashionista’s Choice, which transforms user interaction with fashion by providing intuitive backend systems for real-time engagement. With Fashionista’s Choice, the images of individual pieces of fashion are matched with other items using K-Nearest Neighbors (KNN) algorithms for image retrieval. The visual features are extracted using convolutional neural networks (ResNet50). FastAPI provides the infrastructure for deploying the models yielding low latency responses when users upload their images. Users are authenticated using Firebase Authentication, while Firestore manages real-time access to the metadata of user preferences, product, and category filters. Image uploads are stored within Firebase, and the UI was created with Flutter which also controls user interaction. Fashionista’s Choice provides real-time recommendations tailored to the user, their style intent, and other external factors through an integrated system. Results that are anticipated include shorter search duration, better fashion coordination results, and enhanced interaction with the shop. Real time “Just For You” feeds respond to user actions while visual matching allows users to find items that resemble the styles they adore. Interaction data also helps with trend analysis and targeted product offerings, which is valuable for retailers. To conclude, Fashionista’s Choice AI app eliminates the hassle that comes with searching for fashion and provides users with an ultra-personalized experience. Through handling real-time data and AI-driven image recognition, the application intuitively identifies its users and their identities, changes dynamically to suit the user’s style, and adapts to current fashion trends—offering more than mere product recommendations but rather serving as a smart assistant for a user that is always on-trend.
