Multispectral Photoacoustic Tomography Based On Virtual Parallel-Projection Model-Based Reconstruction Algorithm

Photoacoustic imaging(PAI)is a newly-developing medical image technology that combines the advantages of optical imaging with and ultrasound imaging.It is one of the most rapidly developing technologies in the medical image fields.Photoacoustic tomography(PAT)is one of the most important application area of PAI technology.It is applied to clinical medical imaging and small animals in vivo experiments.Multispectral PAT technology can further realize functional imaging,therefore the PAT has great research and application value.After research and analysis of the existing PAT imaging system.In this paper,a multi-spectral photoacoustic tomography system with multi-angle film illumination mode has been built.The system uses multi-angle sheet light technology to illuminate the imaging objectives.The photoacoustic signals acquired by a detector with the focusing characteristics.In this way,the cross-section imaging capability of the system could be improved effectively.Reconstruction algorithm is important component of imaging technology.The model-based(MB)reconstruction algorithm present quantification accuracies compared with back-projection algorithm.However,the MB method needs to establish a forward model which has a high computational cost in the reconstruction process.The computational cost will increase for large active surface detectors because the geometry area of the detector must be considered.Cylindrically focused transducers with large active surface are widely used in photoacoustic imaging because they provide a high-cross sectional resolution.For the PAT system which employ the large active surface detector,herein,we present a modified model-based algorithm using a virtual parallel-projection method to reduce the computational cost.By this way,the sum of the surface elements' model matrixes can be replaced by a virtual parallel-projection model matrix,the run-time and memory consumption decreased accordingly as well.In this paper,the high resolution characteristic of imaging system has been validated by experiments.The proposed reconstruction method has been validated by the simulations,phantom experiments and in vivo experiments respectively,which exhibits desirable performances in image fidelity.Meanwhile,the forward model which capable of accounting for the shape transducers built by the proposed method without the calculation time and memory requirement increase.Finally,the structural and functional information of small animal are obtained by in vivo experiments.