Assistant professor of biomedical engineering Yu Gan was recently awarded $300,000 from the USDA National Institute of Food and Agriculture for his project “DSFAS-AI: Food Quality Evaluation Leveraging Robust, Domain Adaptive Deep Learning on Millimeter Wave (mmWave) Images.”
In this project, Yu will develop an inexpensive, intelligent imaging system to collect massive food data for quality evaluation; and develop robust learning algorithms to address bias and domain adaptation issues in data science for food quality evaluation. The work takes advantage of artificial intelligence to address unmet data science questions for food quality evaluation. The project is based on an inexpensive imaging system using millimeter-wave radar chipsets.
With this research, Gan hopes to provide a design of bias-robust learning, domain-adaptive, and inexpensive data acquisition, which will ensure that the analysis of massive food data is efficient and effective.
“This project seeks an effective approach to improve utilizing data for food quality evaluation,” said Gan. “My lab is working toward developing cutting-edge techniques to answer two questions in food data science: how to robustly use food images to train computers for a single food evaluation task, and how to train a computer to handle a new food evaluation task when food images are less than expected."
He continued, “Beyond data science, we are working with a hardware engineer, Dr. Nathan Jeong, and a food scientist, Professor Lingyan Kong, both from the University of Alabama, to build an integrated system to test our new ideas in food data science from end to end. We are excited about and thankful for the support from the USDA. We are also initiating collaboration with industry companies to further drive applications of our technique and system in the food science industry.”