Clothing classification using cnn and shopping mall search system

Abstract

This study proposes a system that provides image-based clothing information services in the fashion online market environment among various service industries. This system learns data from CNN and creates a clothing classification system through SAREKnet and VGGNet model. Then, the color was found through k-means and the results were designed to be searched in the shopping mall. Unlike previous studies, we propose the method of using both image search and text search, to increase accuracy. In the future, it is necessary to study through the method of searching for whole-outfit, the method of deriving more class, and the algorithm that can automatically find clothes. © 2020, ICIC International. All rights reserved.