Research On Multi-Constrain Iterative Reconstruction Methods For Photoacoustic Tomography

Photoacoustic tomography(PAT)is a new hybrid imaging mode,which combines the advantages of optical imaging and acoustic imaging with high resolution of ultrasound and high contrast of light.PAT is used for non-invasive structural,functional and molecular imaging in vivo,and has broad application prospect in the early detection of cancer.The main reasons for the low image quality of PAT structure are these,on the one hand,in the forward propagation,nonlinear attenuation occurs when light travels through tissues,leading to the weakening of deep tissue signals.On the other hand,in the acoustic inversion process,due to the ideal description of sound wave propagation and detection,there is an error between the acoustic signal predicted by the model and the real value,resulting in the interference of streak artifacts in the reconstructed image.Therefore,to improve the quality of PAT image to be reconstructed,appropriate penalties or prior knowledge should be incorporated into the imaging model.This paper focuses on the method of improving PAT structure image quality,multispectral PAT imaging technology and the improvements to the imaging technology:(1)Joint constraint Photoacoustic tomography based on the non-local and the sparsity regularizations.The model-based Photoacoustic tomography image reconstruction is an ill-conditioned inverse problem affected by factors such as limited detection angle,imperfect model matrix,and noise.Accounting for this,appropriate penalties should be incorporated into the reconstruction process to improve image quality.In this paper,we present a new dual-constraint PAT imaging model involving a combination of non-local means filtering and sparse coding,with the former to preserve image details by self-similarity and the latter to enforce sparsity.A two-step optimization algorithm and an iterative parameter tuning method were proposed to ensure accurate solution.By comparing to other existing regularization approaches in both numerical simulation and in vivo animal imaging studies,the new method showed improved image quality in terms of signal to noise ratio and contrast enhancement.(2)Multispectral Photoacoustic tomography based on the interlaced sparse sampling.Multispectral photoacoustic tomography(PAT)is capable of resolving tissue chromophore distribution based on spectral un-mixing.It works by identifying the absorption spectrum variations from a sequence of photoacoustic images acquired at multiple illumination wavelengths.Due to multispectral acquisition,this inevitably creates a large dataset.To cut down the data volume,sparse sampling methods that reduce the number of detectors have been developed.However,image reconstruction of sparse sampling PAT is challenging because of insufficient angular coverage.To solve this problem,we present the interlaced sparse sampling(ISS)PAT,a method that involved:1)a novel scanning-based image acquisition scheme in which the sparse detector array rotates while switching illumination wavelength,such that a dense angular coverage could be achieved by using only a few detectors;and 2)a corresponding image reconstruction algorithm that makes use of an anatomical prior image created from the ISS strategy to guide PAT image computation.Simulation,phantom,and in vivo animal experiments showed that our method achieved comparable image reconstruction and spectral un-mixing results to those obtained by conventional dense sampling method.(3)Multispectral interlaced sparse sampling Photoacoustic tomography with directional total variation.Multispectral interlaced sparse sampling photoacoustic tomography(ISS-PAT)is a novel scanning-based image acquisition scheme.And in the existing research on ISS-PAT,the prior information is introduced by adopting the non-local mean(NLM)method that can well preserve local structure.However,since NLM based method needs to find similar patches across the whole image,the computational burden is huge.To improve the speed of reconstruction,a new method named directional total variation(dTV)is used to induce the structural information of the combined image.We solve the minimization problem with the alternating direction method of multipliers(ADMM).Qualitative and quantitative evaluations are performed on simulation,phantom,and in vivo animal experiments.Experimental results show that dTV method is very effective and reconstruction speed is improved more than 10 times compared with the NLM method.Furthermore,the image quality of dTV is comparable to that of NLM.