Model-Based Imaging: An Integration of Physics, Signal Processing, and Biology

A classical problem in biomedical imaging is the so-called limited data problem, which occurs when physical and temporal constraints prevent sufficient coverage of the data space. Traditionally, image reconstruction is formulated as a linear inverse problem, which is usually solved using Fourier transform-based methods. This approach often leads to a significant loss of spatial/temporal resolution, which limits the practical utility of various imaging modalities. To overcome these problems, numerous methods have emerged in the past two decades to incorporate a priori information into the imaging process. This talk will provide a systematic discussion of model-based imaging. I will begin with a brief review of the diffraction-limited imaging problem and then discuss in detail how a priori information can be effectively utilized to achieve high spatial and temporal resolution in various biomedical applications (e.g., cardiac imaging, functional neuroimaging, and molecular cancer imaging).