Medical Engineering Defense, Xin Tong

Zoom link: https://caltech.zoom.us/j/89859626512

Optical imaging enables visualization of biological structure and function with remarkable versatility, yet it is fundamentally limited by several physical constraints. Spatial resolution is bounded by optical diffraction, depth penetration is curtailed by strong absorption and scattering in tissue, and image contrast-to-noise ratio is often restricted by photon shot noise in low-light conditions. Two novel directions—photoacoustic imaging and quantum imaging—seek to address these limitations through distinct physical mechanisms. During my Ph.D. training, I aim to develop new technologies in both domains to advance biomedical imaging.

In photoacoustic imaging, I design and optimize high-speed photoacoustic computed tomography systems that enable deep, volumetric visualization of vasculature. By incorporating time-gated reconstruction and image-enhancement algorithms, these systems support small-animal imaging of cardiac structure, liver morphology, and hemodynamics under both physiological and perturbed conditions. Building on these foundations, I explore non-invasive breast photoacoustic imaging with high spatiotemporal resolution. Through integration with classical and learning-based feature extraction, classification, and segmentation pipelines, I demonstrate the feasibility of applying photoacoustic imaging in clinical workflows to aid the assessment and characterization of breast tissue.

In quantum imaging, I characterize entangled photon-pair sources in spatial and polarization domains to understand how these degrees of freedom enhance remote sensing and image formation. Leveraging these insights, I develop a large-field-of-view scanning quantum imaging platform that combines single-photon counting hardware with coincidence-based detection, achieving the first demonstration of whole-organism imaging using entangled photons. I further engineer a quantum microscopy system that pairs an electron-multiplying charge-coupled device with a covariance-based coincidence estimation algorithm, achieving two-fold resolution beyond the classical diffraction limit for cellular imaging.

Across both research directions, this thesis advances system design and engineering, quantitative characterization, calibration, reconstruction, and image-enhancement methodologies. Together, these developments establish pathways from physical principles to practical imaging systems, spanning laboratory prototypes through preclinical and clinical applications in biological and medical imaging.