Fashion Classification and Recommendation System Based on User Body Shape Using Convolutional Neural Network and Content Based Filtering Methods

Consumers currently have unlimited access to a variety of online products and services, including clothing. Despite significant progress, online clothing shopping often has challenges, especially in choosing clothing designs that match the user's body shape type, especially for women to support their appearance. Choosing the right clothes can increase the wearer's confidence and comfort. Therefore, it is necessary to create a tool that can be used by the public in choosing the right clothes. This study aims to design and implement a clothing classification and recommendation system that combines artificial intelligence and content-based recommendation techniques. The CNN method is used to classify the user's body shape type based on the images taken and then the system will provide clothing recommendations that match the user's body shape type using the Content-Based Filtering method. This system can integrate CNN methods very well to classify the user's body shape type and then recommend products that match the classified body shape type using the CBF method. Dataset scenario with 90% training data, 10% testing data and learning rate of 0.001 produces very good accuracy with training results of 96.61% and testing results of 85.71%. This system can provide a more satisfying shopping experience, providing assurance that each clothing choice is not only fashionable but also in line with the user's visual identity and preferences.