Consumer satisfaction is THE quality metric for FLASH; however, it is very challenging to evaluate consumer satisfaction remotely. Students on the FLASH team will design and develop a computer vision system for FLASH parking kiosks to determine consumer sentiment (satisfied/dissatisfied) in an anonymized and balanced manner, and then utilize the trend in sentiment to help discover operational problems.
Abstract:
Facial Expression Recognition (FER) is the task of classifying the expressions on face images into various categories, such as anger, fear, surprise, happiness, etc. FER is moving from laboratory-based theory demonstrations to practical real- world applications, but there still many challenges to achieving a working system in the field. Automatic FER has typically faced two major challenges: lack of sufficient and/or demographically representative training data, and variation in image quality parameters. Although there are 6 or 7 sentiments that are frequently identified, in order to deliver a benefit to FLASH, this system only needs to categorize into satisfied (good/neutral experience) or dissatisfied (poor/bad experience). The team will utilize a historical dataset of images to develop the recognition system, and correlate the trend in consumer sentiment to operational functioning within our parking garages. Validation will include field testing of the recognition engine.
Impact:
A working system will provide a new source of information to understand consumer satisfaction, provide a mechanism to evaluate the effectiveness of design changes in the product, and even help to improve operational responsiveness to problems.
Computer Vision (2-4 Students)
Specific Skills: Image processing, Computer Vision and Machine Learning, Algorithm and tool development.
EECS 281 (or equivalent) is required.
Experience with or willingness to learn Python.
Likely Majors: CS, ROB, DATA, CE, EE
Software Development and Database Design (2-3 Students)
Specific Skills: Database design for images and attribute solid understanding of database design theory and practical knowledge, General Programming.
Completion of EECS 281 or equivalent is required.
Prior experience with AWS is a plus.
Likely Majors: CS, DATA
Statistical Analysis (1 Student)
Specific Skills: Statistics of trend identification.
Design of Experiments (for field validation).
Likely Majors: STATS, DATA
Sponsor Mentor
Hunter Dunbar
Hunter Dunbar manages AI/Computer Vision at FLASH. He’s held previous positions in engineering and product at venture backed companies, deploying Machine Learning products in commercial buildings and fleet vehicles. In his private time, he enjoys tinkering at the intersection of furniture and robotics, as well as kiteboarding. He received his collegiate education at the University of Minnesota.
Executive Mentor
Jeff Judge
Jeff is CTO at FLASH. As CTO, he is responsible for the product, engineering, and data teams efforts at the company, and ensuring that these teams work together as a unified front. Prior to FLASH, Jeff was cofounder of two companies, Bright and Signal, and an early engineer at Orbitz. Jeff is an active angel investor and Techstars mentor. Jeff lives in the Lincoln Square neighborhood of Chicago with his wife and five children.
Faculty Mentor
Kayvan Najarian, Ph.D.
The focus of Dr. Najarian’s research is on the design of signal/image processing and machine learning methods to create computer-assisted clinical decision support systems that improve patient care and reduce the costs of healthcare. Dr. Najarian’s lab also designs sensors to collect and analyze physiological signals and images, focusing on creating decision support systems to manage traumatic brain injuries, traumatic pelvic/abdominal injuries and hypovolemia. He serves as the Editor-in-Chief of Biomedical Engineering and Computational Biology and the Associate Editor of two other journals in the field of biomedical informatics.
Citizenship Requirements:
- This project is open to all students on campus.
- International students on an F-1 visa will be required to declare part-time CPT during Winter 2023 and Fall 2023 terms.
IP/NDA: Students will sign standard University of Michigan IP/NDA documents.
Internship/Summer Opportunity: Students will be guaranteed an interview for a 2023 internship. The interviews will take place in Jan/Feb 2023.