Research

Research Overview

Vision: transform the biomedical technology, clinical practice, and precision healthcare by developing and applying innovative artificial intelligence and other computational methods for building biomedical image-centered systems with high precision and automacy to deliver better healthcare. See my Research Statement (full, short). 

Expertise: machine/deep learning, computer vision and medical image analysis, convex and non-convex optimization, and statistical signal and image processing. 

Past research: adversarial robustness of deep learning systems, compressed sensing, large scale convex and non-convex optimization, information-theoretical foundation of deep learning, computer vision, diagnostic breast image analysis, multimodality-based cardiovascular disease diagnostics and prognostics, and cancer treatment planning. 

Research Interests: I'm interested in biomedical imaging and image analytics such as efficient and principled algorithms design, and their applications in biomedicine and healthcare. More specifically, my recent interests and efforts span the following topics 

• biomedical imaging and image analysis such as CT/SPECT/PET/MRI/US-based image reconstruction, super-resolution, classification, object detection, segmentation, generative AI, multimodality large language model (MLLM), agent AI. More 

• AI in medicine and precision health such as CT/SPECT/PET/MRI/US-based quantification and biomarker discovery, risk stratification, diagnosis, and treatment planning in breast cancer imaging and cardiovascular imaging. More 

• machine learning and deep learning such as generalizable learning, accelerated training optimization algorithms, explainable AI algorithms design, and learning with small sample size. More 

A complete list of publications can be found at Google Scholar

Seminars and Invited Talks

• A progressive training approach for object detection in computer aided diagnosis. R&D AI Seminar, Hologic Inc, Santa Clara, 01/24/2023.

• Robust AI and domain adaptation series - part I. R&D AI Seminar, Hologic Inc, Santa Clara, California, USA, 11/16/2022.

• Vision transformer series - part I. R&D AI Seminar, Hologic Inc, Santa Clara, California, USA, 02/01/2022 and 02/09/2022.

• Separation-free super-resolution from compressed measurements is possible: an orthonormal atomic norm minimization approach. IEEE International Symposium on Information Theory, June, 2018.

Media Coverage

• Electrical and Computer Engineering professors receive National Science Foundation RAPID grant for novel high-throughput and low-cost COVID-19 testing technologies, University of Iowa College of Engineering News. link

• Pooling samples could help you get your COVID-19 test results faster, IEEE Spectrum. link

• The mathematics of mass testing for COVID-19, SIAM News. link

• Advances rapid COVID-19 testing, University of Iowa Dare to Discover Campaign. link