Smart Wardrobe: A Comprehensive Approach to Personalized Clothing Recommendation with The Use of Nearest Neighbor Model
Main Article Content
Abstract
In the era of digital assistance, selecting the perfect outfit for various occasions can be overwhelming. To alleviate this challenge, this research proposes a multi-model approach for personalized clothing recommendations. Our system integrates rule-based classification, deep learning with ResNet-50, and K Nearest Neighbor (KNN) algorithms to provide tailored outfit suggestions to users. Firstly, a rule-based classifier maps user-specified occasions to clothing categories (formal, traditional, casual). Next, a fine-tuned ResNet-50 model analyzes uploaded outfit photos to predict suitable attire. Subsequently, KNN is employed to recommend outfits based on either similarity to a seed image or items within the user's wardrobe. The system generates outfit recommendations, encompassing both clothing categories and specific combinations, enhancing user convenience. Furthermore, the integration of e-commerce links enables seamless access to purchase recommended attire, ensuring a comprehensive user experience. Our proposed system offers a holistic solution to the challenge of outfit selection, leveraging the strengths of machine learning and Rule-Based logic to provide personalized and actionable recommendations.
Article Details
This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.