NCSI: Large Scale Clothing Retrieval Using Deep Fashion Landmarks

Status	Completed
Mentor Name	Hao Ji
Mentor's XSEDE Affiliation	Education Allocation
Mentor Has Been in XSEDE Community	4-5 years
Project Title	Large Scale Clothing Retrieval Using Deep Fashion Landmarks
Summary	Recent advances in machine learning, especially in deep learning, have made a big leap forward, leading to state-of-the-art fashion recommender systems with both high accuracy and fast speed. In this project, we plan to build a large-scale recommender system for retrieving fashion items of interest by taking advantage of deep fashion landmarks. We will compare with the approaches based on handcrafted vision features and/or deep CNN features in terms of accuracy and speed.
Job Description	The student will develop web crawlers to collect fashion items from fashion websites. Those fashion images will be indexed based on deep fashion landmarks obtained by deep learning models for fast search. The student will benchmark the recommender system with "in the wild" fashion datasets, in comparisons with handcrafted vision features and/or deep CNN features.
Computational Resources	XSEDE Bridges' GPU clusters
Contribution to Community
Position Type	Intern
Training Plan	I will offer a "Deep Learning" course in the spring semester of 2019. This course will further students' knowledge and skills in Machine Learning. During this project, I will work closely with the students and plan to have weekly meetings with them to guide their research progress.
Student Prerequisites/Conditions/Qualifications	Must be an undergraduate at California State Polytechnic University, Pomona Must have a good understanding of machine learning, linear algebra, and good programming skills in Python.
Duration	Semester
Start Date	01/19/2019
End Date	05/10/2019