I-Corps: Artificial intelligence-based image processing to automate fashion e-commerce

点击翻译按钮获取中文摘要

The broader impact/commercial potential of this I-Corps project is the development of software tools to automate several essential tasks for internet-based commerce (e-commerce). E-commerce is a rapidly growing market segment, predicted to reach $6.5 trillion in sales by 2023. Consequent to such rapid growth are the resultant challenges in product discovery. Rich product attributes/product tagging makes products easily discoverable. However, populating product attributes/tags is a time-consuming, mostly manual task. By leveraging artificial intelligence (AI)-based image processing, these tasks may be automated, saving time, effort, and expense for e-commerce stores. In addition, accurate tags may be used to generate several useful functionalities to improve user experience as well as increase sales and revenue for e-commerce stores. AI-based automation tools may provide e-commerce store owners with access to advanced technologies and empower small and mid-sized businesses to stay viable in the highly competitive fashion e-commerce market segment. This I-Corps project is based on the development of image processing technology that uses machine learning and artificial intelligence (AI) to accurately recognize multiple (50+) attributes associated with an image. These attributes are used to automatically and accurately tag products in the store at a fraction of a cost and time compared to manual tagging. Accurate tagging of products results in making products easily discoverable online. The AI-algorithm generated tags also are used to provide several additional functionalities such as complete-the-look and product bundling. Such functionalities significantly improve the user experience for shoppers and lead to increased sales and revenue for store owners. The proposed technology also includes resource optimization to make AI-powered processing commercially viable. Such resource optimization ensures that advanced AI-based technologies are available for dynamic environments such as fashion e-commerce at reasonable costs. In addition, the process flow between the AI-algorithm, machine learning-based decision tree, and the output ensures that the results are displayed at the storefront almost instantaneously.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.