Jirong Yi

Welcome to Jirong Yi's webpage.

Short bibliography

I received a PhD degree in computer engineering from University of Iowa under supervision of Prof. Weiyu Xu and Xiaodong Wu. My research has been centered around computational medical imaging-based precision health. I have long-standing and close collaboration with clinicians and industrial partners including cardiologists, radiologists, medical physicists, radiation oncologists, and medical imaging AI experts.

My work at University of Iowa focused on (1) compressed sensing and deep learning methods and theory for medical image reconstruction and quality enhancement, and (2) large-scale convex optimization for CT-guided cancer radiation therapy treatment planning optimization. My work has been published in leading venues in the fields such as IEEE Transactions on Information Theory (T-IT), OUP Information and Inference: a Journal of the IMA, International Conference on Learning Representations (ICLR). My work has also been translated into patents. See my thesis and video.

During my time with the Computer-aided Diagnosis (CAD) Science team at Hologic Inc, I worked on (1) deep learning for mammogram- and tomosynthesis-based image classification and image object detection, and (2) their applications in breast cancer diagnostic product development and deployment. My work has been deployed into Hologic breast cancer diagnostic imaging product stack, cleared by the Food and Drug Administration (FDA), and dominating the breast imaging market.

Since 2023, I have been with Prof. Piotr Slomka's team at Cedars-Sinai Medical Center, and working on (1) deep learning for cardiac perfusion SPECT/CT and PET/CT multimodality image analysis; (2) their applications in quantitative medical imaging biomarker discovery, and cardiovascular disease prevention, diagnosis, and treatment; (3) clinical validation of medical imaging AI tools in retrospective cohort studies with over 73K patients; (4) construction of the world largest nuclear multimodality medical imaging (spanning PET, SPECT, CT, eletronical medical records, text reports, clinical data) databases of over 46K patients. My work has been deployed to process the SPECT/CT, PET/CT scans and electronic medical records data of over 46K patients, and been published in leading venues in the fields such as Lancet Digital Health, European Heart Journal, npj Digital Medicine, and Cardiovascular Research.

In summary, my research has been producing publications in leading journals (e.g., Lancet Digital Health), FDA-approved products (e.g., Genius AI Detection for tomosynthesis-based breast cancer diagnosis), and patents (e.g., CT-guided cancer radiation therapy treatment planning). See Google Scholar, and Linkedin for more.

Education

Ph.D., Computer Engineering, University of Iowa, Iowa City, Iowa, USA, 2021
M.S., Computer Engineering, University of Iowa, Iowa City, Iowa, USA, 2020
B.S., Information and Computing Science, Harbin Engineering University, Harbin, Heilongjiang, PRC, 2015

Experience

Postdoctoral Scientist, Division of Artificial Intelligence in Medicine, Cedars-Sinai Medical Center, Los Angeles, California, USA, 10/2023-Now
Machine Learning Research Scientist, Computer-Aided Diagnosis Science, Hologic Inc, Santa Clara, California, USA, 09/2021-08/2023
Engineer, Department of Electrical and Electronic Engineering, Southern University of Science and Technology, Shenzhen, Guangdong, China, 07/2015-07/2016

Interests

My interest lies in the intersection of AI/ML, medical image analysis and informatics, and their translation into clinical practice and precision medicine, specifically in the following highly aligned topics:

Medical image understanding and informatics

including compressed sensing, computer vision, machine learning and deep learning methods for medical image classification, image object detection, image segmentation, image registration, generative imaging AI, multi-modality image fusion and alignment, vision-language foundation models, agentic imaging AI systems.

Translation into precision medicine

including machine learning, computer vision, machine learning and deep learning techniques for mdical imaging biomarker discovery and quantification, and disease prognosis, diagnosis, and treatment in cardiovascular disease and cancer.

AI and computer vision methodologies

including imbalanced learning, fine-grained learning, small-sample learning, explainable AI, domain adaptation, self-supervised learning, semi-supervised learning, and model calibration.

Vision

My long-term goal is to accelerate innovation and clinical translation in biomedical imaging AI by strategically aligning cutting-edge AI technology development, advanced medical image analysis, real-world unmet clinical needs, and scalable solutions.