Research On Compensation Method Of Acoustic Reflection And Attenuation In Endoscopic Photoacoustic Tomography

Endoscopic photoacoustic tomography(EPAT)provides reliable information for the accurate diagnosis of cavity diseases and guidance of interventional treatment.Due to the acoustic inhomogeneity of the imaged tissue,photoacoustically generated ultrasound will be scattered,reflected and absorbed at the boundaries of different tissue layers.Consequently,in image reconstruction,the assumption of ideal acoustic properties usually results in the reduction in the image quality and imaging depth.The purpose of this thesis is to design methods for reconstructing high-quality EPAT images in acoustically heterogeneous tissues in order to reduce artifact,blurring and distortion in the images caused by acoustic reflection and attenuation.The main work includes two parts.First,image artifacts due to acoustic reflection are suppressed by retrieving the ideal photoacoustic signals without reflected clutters from the acoustic measurements based on nonlinear least square optimization.Second,the acoustic attenuation is compensated by time reversion based on the acoustic attenuation model to reconstruct the images representing the maps of initial pressure distribution on the cavity cross section from the filtered pressure data by a Tukey time-varying window.The performance of the developed methods have been evaluated with computer-simulated phantoms of coronary arterial vessels and general cavities containing diseased tissues.Comparison experiments are also conducted comparing the reconstructions provided by our methods,conventional reconstruction approaches,short lag spatial coherence method,deep learning approach and singular value decomposition method to verify the superiority of our methods in reconstructing high-quality images.The results demonstrate the improvement in the contrast and quality of images reconstructed with our methods.