Research on Image Reconstruction Algorithm for Intravascular Photoacoustic Imaging

Intravascular photoacoustic (IVPA) imaging is a novel intravascular endoscopic imaging method to circumvent the limitations of existing imaging modalities. It directly images morphological and pathological changes in vessel walls, as well as the contents of vulnerable plaques, with promising clinical applications in vascular disease diagnosis and therapy. In this dissertation, the imaging mechanism of IVPA and physical properties of the photoacoustic signals of vessel wall tissues are examined. The deconvolution problem, specifically the detector’s pulse response effect, is tackled with a regularization method to effectively recover the photoacoustic signals. Image reconstruction from the recovered signals is crucial to IVPA imaging, and its quality depends significantly on the reconstruction algorithm. The thesis summarizes current reconstruction algorithms for in vitro photoacoustic imaging and presents two reconstruction algorithms for IVPA imaging using a single-element ultrasonic transducer: filtered back-projection and time-reversal algorithms. The algorithms are applied to a computer-simulated vessel phantom, and the results comparison is presented. Sparse and limited-angle sampling effects on image reconstruction are discussed. The observation is that only then the tissue boundaries may be reconstructed accurately when their normal lines pass through the sampling position of the detector.