Deep learning for image reconstruction in thermoacoustic tomography

Microwave-induced thermoacoustic tomography(TAT)is a rapidly-developing noninvasive imaging technique that integrates the advantages of microwave imaging and ultrasound imaging.While an image reconstruction algorithm is critical for the TAT,current reconstruction methods often creates significant artifacts and are computationally costly.In this work,we propose a deep learning-based end-to-end image reconstruction method to achieve the direct reconstruction from the sinogram data to the initial pressure density image.We design a new network architecture TAT-Net to transfer the sinogram domain to the image domain with high accuracy.For the scenarios where realistic training data are scarce or unavailable,we use the finite element method(FEM)to generate synthetic data where the domain gap between the synthetic and realistic data is resolved through the signal processing method.The TAT-Net trained with synthetic data is evaluated through both simulations and phantom experiments and achieves competitive performance in artifact removal and robustness.Compared with other state-of-the-art reconstruction methods,the TAT-Net method can reduce the root mean square error to 0.0143,and increase the structure similarity and peak signal-to-noise ratio to 0.988 and 38.64,respectively.The results obtained indicate that the TAT-Net has great potential applications in improving image reconstruction quality and fast quantitative reconstruction.     

The adaptive sinogram restoration algorithm based on anisotropic diffusion by energy minimization for low-dose X-ray CT

Apparent streak artifacts will present in reconstructed images due to excessive quantum noise in low-dose X-ray imaging process. Estimating a noise-free sinogram to satisfy the filtered back-projection (FBP) reconstruction is an effective way to solve this problem. In this paper, we propose a novel sinogram noise reduction method by energy minimization. An adaptive smoothness parameter based on a modified anisotropic diffusion coefficient is applied for an optimal estimation. The smoothness parameter can make the method effectively adjust the degree of smoothness according to the noise level and the region feature in the sinogram. Visual effect together with quantitative analysis of the experimental result shows the developed approach has the excellent performance in protection of the edge and removal of streak artifacts in the reconstructed image.     

A statistical iteration approach with energy minimization to sinogram noise reduction for low-dose X-ray CT

Low-dose CT imaging has been particularly used in modern medical practice for its advantage on reducing the radiation dose to patients. However, excessive quantum noise is present in low dose X-ray imaging along with the decrease of the radiation dose; thus, there are obvious streak-like artifacts in reconstructed images. The statistical iterative reconstruction approach applied to the noisy sinogram before a filtered back-projection (FBP) is a resolution to deal with the noisy problem. In this paper, the statistical property of the noise sinogram was considered to achieve a satisfactory image reconstruction and a statistical iterative method with energy minimization was proposed to address the problem of streak-like artifacts. Simulations were performed and indicated that the proposed method could suppress noise and obviously decrease streak-like artifacts in reconstructed images.     

基于加权TV正则化的X射线CT系统能谱估计方法

在CT硬化伪影校正、双能CT图像重建以及CT辐射剂量计算等实际应用中,X射线的能谱信息具有重要的作用。然而,由于透射测量方程组系数矩阵的病态性、X光子的统计涨落和噪声的干扰,使得EM等能谱估计方法难以获得较精细的能谱刻画。针对该问题,提出一种基于加权TV正则化的X射线CT系统能谱估计方法。首先采用能谱能量范围内带有不同K-edge的材料作为体模以降低投影测量方程之间的相关性。然后,利用CT成像系统的几何参数来获取投影测量数据所对应的准确透射长度信息来减小投影方程的测量误差。最后,综合利用透射衰减测量数据的保真性、轫致辐射能谱部分的连续性、特征辐射能谱部分的离散性、能谱的非负性和归一性以及平均有效衰减系数等信息,使用加权TV正则化方法建立目标函数,利用正则化理论中的L曲线准则通过黄金分割参数搜索策略求得最优正则化参数及对应的能谱估计结果。分别通过不同统计波动模型下的仿真能谱和实际测量数据对算法性能进行了验证,结果表明该方法与EM等方法相比,不受初始能谱的影响,有效提高了能谱估计结果的稳定性和准确性。     

Improved simultaneous multislice magnetic resonance imaging using total variation regularization

Controlled aliasing in parallel imaging results in higher acceleration(CAIPIRINHA) for simultaneous multislice imaging has been proposed recently, which combines multiband excitation and phase cycling techniques to reduce scan time and improve subsequent imaging reconstruction. In this work, the total variation(TV) regularization method is used to further improve CAIPIRINHA. The TV regularization uses an edge-preserving prior, which establishes a relationship between neighboring pixels for image reconstruction. It reduces artifacts and suppresses noise amplification simultaneously.The results are presented with a standard eight-channel head coil with an acceleration factor of 4, where the TV-regularized CAIPIRINHA generates an improved reconstruction as compared with a typical nonregularized CAIPIRINHA.     

基于极坐标变换去除计算机层析图像环形伪影

探测器像元响应的不一致、光源能不稳定等因素使得计算机层析(CT)图像中含有较多的环形伪影,严重降低了图像质量,影响了图像的三维重建和量化分析,为此提出了基于极坐标变换与傅里叶变换后低通滤波的算法去除环形伪影。通过极坐标变换将直角坐标下的环形伪影转化为极坐标下的线性伪影,然后对线性伪影图像进行傅里叶变换获得频谱图像,进而设计二维低通滤波器进行滤波处理,最后通过傅里叶逆变换与坐标逆变换获得校正后的图像。利用Matlab软件,编写程序对算法进行验证,结果表明,该算法能够有效地去除环形伪影,使图像内部细节清晰可见,并且保护了图像边缘信息,提高了图像的信噪比;另外,使用该方法处理100张切片图像只需3.5min,可满足批量处理的需求。     

Noise reduction for low-dose X-ray CT based on fuzzy logical in stationary wavelet domain

Apparent streak-like artifacts will present in reconstructed images due to excessive quantum noise in low-dose X-ray imaging process. Dealing with the noisy sinogram before a filtered back-projection (FBP) is a useful solution to solve this noise problem. In this paper, we proposed a novel noise restoration method combining wavelet and fuzzy logical technology for low-dose computed tomography (CT) sinogram. The method first utilizes stationary wavelet transform on the noisy sinogram, namely decomposes the sinogram to different resolution levels. And then, at each decomposed resolution level, a fuzzy shrinkage filter is applied to restore the noise-contaminated wavelet coefficients. Simulations were performed and indicated that the proposed method could significantly suppress noise and reduced streak-like artifacts in reconstructed images while at the same time maintaining the image sharpness.     

Reduced aliasing artifacts using shaking projection k-space sampling trajectory

Radial imaging techniques, such as projection-reconstruction （PR）, are used in magnetic resonance imaging （MRI） for dynamic imaging, angiography, and short-T2 imaging. They are less sensitive to flow and motion artifacts, and support fast imaging with short echo times. However, aliasing and streaking artifacts are two main sources which degrade radial imaging quality. For a given fixed number of k-space projections, data distributions along radial and angular directions will influence the level of aliasing and streaking artifacts. Conventional radial k-space sampling trajectory introduces an aliasing artifact at the first principal ring of point spread function （PSF）. In this paper, a shaking projection （SP） k-space sampling trajectory was proposed to reduce aliasing artifacts in MR images. SP sampling trajectory shifts the projection alternately along the k-space center, which separates k-space data in the azimuthal direction. Simulations based on conventional and SP sampling trajectories were compared with the same number projections. A significant reduction of aliasing artifacts was observed using the SP sampling trajectory. These two trajectories were also compared with different sampling frequencies. ASP trajectory has the same aliasing character when using half sampling frequency （or half data） for reconstruction. SNR comparisons with different white noise levels show that these two trajectories have the same SNR character. In conclusion, the SP trajectory can reduce the aliasing artifact without decreasing SNR and also provide a way for undersampling recon- struction. Furthermore, this method can be applied to three-dimensional （3D） hybrid or spherical radial k-space sampling for a more efficient reduction of aliasing artifacts.     

Ring artifacts correction based on the projection-field in neutron CT

Ring artifacts will happen mostly when the detector has inconsistent response among the detector channels, and the characteristic produced rings centered in the iso-center in the reconstructed slices inevitably affect the recognition and analysis of the corresponding sample structures in neutron computed tomography(CT). In this work, a ring correction method based on the projection-field(RCP) is proposed, it is a pre-processing method and provides the corrected projection data directly, which is also conducive to efficient data storage and other algorithmic researches. Simulation and physical experiments are performed for verifying the effect of the method, and one of the correction methods based on the imagefield is used for comparison. The results demonstrate that the RCP can correct the ring artifacts well without reducing the image resolution or over-correction.     

Estimation of mutual information objective function based on Fourier shift theorem: an application to eddy current distortion correction in diffusion tensor imaging

Diffusion tensor imaging requires correction of eddy current distortion in diffusion-weighted images. An effective retrospective correction approach is to transform a diffusion-weighted image to maximize the mutual information (MI) between the transformed diffusion-weighted image and the corresponding T2-weighted image. In the literature, either linear interpolation or partial volume interpolation is applied to estimate the MI objective function. However, these interpolation methods induce artifacts to the MI objective function, thus compromising correction results. In this work, the MI objective function is estimated based on interpolation using Fourier shift theorem. This method eliminates the artifacts incurred with the aforementioned interpolation methods. The algorithm is further improved by approximating pixel values using their nearest neighbors in the up-sampled spatial domain, resulting in dramatically increased computational efficiency without compromising the correction results. The effects of varying the number of quantization levels and using Parzen window filtering to smooth the MI objective function are also investigated to obtain optimized algorithm parameters. The diffusion tensor image quality after applying the proposed distortion correction method is significantly improved visually.     

Guided filter and adaptive learning rate based non-uniformity correction algorithm for infrared focal plane array

Imaging non-uniformity of infrared focal plane array (IRFPA) behaves as fixed-pattern noise superimposed on the image, which affects the imaging quality of infrared system seriously. In scene-based non-uniformity correction methods, the drawbacks of ghosting artifacts and image blurring affect the sensitivity of the IRFPA imaging system seriously and decrease the image quality visibly. This paper proposes an improved neural network non-uniformity correction method with adaptive learning rate. On the one hand, using guided filter, the proposed algorithm decreases the effect of ghosting artifacts. On the other hand, due to the inappropriate learning rate is the main reason of image blurring, the proposed algorithm utilizes an adaptive learning rate with a temporal domain factor to eliminate the effect of image blurring. In short, the proposed algorithm combines the merits of the guided filter and the adaptive learning rate. Several real and simulated infrared image sequences are utilized to verify the performance of the proposed algorithm. The experiment results indicate that the proposed algorithm can not only reduce the non-uniformity with less ghosting artifacts but also overcome the problems of image blurring in static areas.     

X射线CT环形伪影去除方法

为去除X射线CT图像中的环形伪影,建立了一种投影数据预处理方法。该方法通过对各角度投影数据逐一进行分段多项式拟合的方式建立异常探元的备选校正因子集合,再根据备选校正因子的概率密度分布确定最可几校正因子。介绍了方法的物理依据、原理及其实现步骤,并分析了其在复杂条件下的适应性。结果表明:该方法对稀疏环形伪影、密集环形伪影以及伴随强噪声污染的CT图像环形伪影均可去除;与中值滤波等方法相比,可以更好地保持图像的空间分辨率。该方法可用于多材质检测对象CT图像的处理。     

数据插值对正电子发射断层成像设备的图像重建影响的研究

正电子发射断层扫描(positron emission computed tomography, PET)是核医学领域最先进的临床检查影像技术. PET技术是目前临床上用于诊断和指导治疗肿瘤的最佳手段之一. 正电子发射断层成像设备探测器采集到的数据需要进行数据处理, 把原始数据转换成正弦图形式的数据才能用于图像重建. 平行束断层重建和扇形束图像重建是图像重建的两种方法, 分别对应平行束和扇形束形式的数据处理方法. 对原始数据的操作不可避免地破坏了原始数据的完整性. 现今, 正电子发射断层设备在重建过程中普遍采用平行束重建的方法. 平行束的数据分离会对PET数据进行插值操作, 扇形束的数据分离不会对PET数据进行插值操作. 本文通过对比平行束图像重建和扇形束图像重建结果, 研究了数据插值对PET图像重建结果的影响. 关键词： 正电子发射计算机断层扫描 数据插值 图像重建 原始数据     

基于深度神经网络的时空编码磁共振成像超分辨率重建方法

单扫描时空编码磁共振成像是一种新型超快速磁共振成像技术,它对磁场不均匀和化学位移伪影有较强的抵抗性,但是其固有的空间分辨率较低,因此通常需要进行超分辨率重建,以在不增加采样点数的情况下提高时空编码磁共振图像的空间分辨率.然而,现有的重建方法存在迭代求解时间长、重建结果有混叠伪影残留等问题.为此,本文提出了一种基于深度神经网络的单扫描时空编码磁共振成像超分辨率重建方法.该方法采用模拟样本训练深度神经网络,再利用训练好的网络模型对实际采样信号进行重建.数值模拟、水模和活体鼠脑的实验结果表明,该方法能快速重建出无残留混叠伪影、纹理信息清楚的超分辨率时空编码磁共振图像.适当增加训练样本数量以及在训练样本中加入适当的随机噪声水平,有助于改善重建效果.     

Cardiac magnetic resonance imaging using radial k-space sampling and self-calibrated partial parallel reconstruction

Radial sampling has been demonstrated to be potentially useful in cardiac magnetic resonance imaging because it is less susceptible to motion than Cartesian sampling. Nevertheless, its capability of imaging acceleration remains limited by undersampling-induced streaking artifacts. In this study, a self-calibrated reconstruction method was developed to suppress streaking artifacts for highly accelerated parallel radial acquisitions in cardiac magnetic resonance imaging. Two- (2D) and three-dimensional (3D) radial k-space data were collected from a phantom and healthy volunteers. Images reconstructed using the proposed method and the conventional regridding method were compared based on statistical analysis on a four-point scale imaging scoring. It was demonstrated that the proposed method can effectively remove undersampling streaking artifacts and significantly improve image quality (P<.05). With the use of the proposed method, image score (1–4, 1=poor, 2=good, 3=very good, 4=excellent) was improved from 2.14 to 3.34 with the use of an undersampling factor of 4 and from 1.09 to 2.5 with the use of an undersampling factor of 8. Our study demonstrates that the proposed reconstruction method is effective for highly accelerated cardiac imaging applications using parallel radial acquisitions without calibration data.     

锥束CT成像质量影响因素研究

CT成像质量受诸多因素影响,有必要系统地研究各因素带来的影响以得到更好的成像效果。基于兰州大学核科学与技术学院研制的锥束CT系统,在近探测器几何条件下,采用一铝制标准件,通过对比实验研究了投影采集范围及步长、管电压及管电流、焦点尺寸、样品在转台位置、硬化校正和图像优化等因素对CT系统成像质量的影响。结果表明,当投影采集完备时,投影采集范围对成像质量影响较小,减小扫描步长能提高成像质量;适当提高管电压能降低硬化伪影,提高管电流能减小图像噪声;较小的焦点尺寸能提高图像空间分辨率,但在近探测器几何条件下不明显;样品在转台位置不影响CT系统还原样品结构;硬化校正能明显消除硬化伪影;最后,对于单一材质样品通过阈值去噪能优化图像质量。以上研究为CT系统的研制和应用提供了参考。     

Fast multi-contrast MRI reconstruction

Multi-contrast magnetic resonance imaging (MRI) is a useful technique to aid clinical diagnosis. This paper proposes an efficient algorithm to jointly reconstruct multiple T1/T2-weighted images of the same anatomical cross section from partially sampled k-space data. The joint reconstruction problem is formulated as minimizing a linear combination of three terms, corresponding to a least squares data fitting, joint total variation (TV) and group wavelet-sparsity regularization. It is rooted in two observations: 1) the variance of image gradients should be similar for the same spatial position across multiple contrasts; 2) the wavelet coefficients of all images from the same anatomical cross section should have similar sparse modes. To efficiently solve this problem, we decompose it into joint TV regularization and group sparsity subproblems, respectively. Finally, the reconstructed image is obtained from the weighted average of solutions from the two subproblems, in an iterative framework. Experiments demonstrate the efficiency and effectiveness of the proposed method compared to existing multi-contrast MRI methods.     

Total variation based gradient descent algorithm for sparse-view photoacoustic image reconstruction

In photoacoustic imaging (PAI), reconstruction from sparse-view sampling data is a remaining challenge in the cases of fast or real-time imaging. In this paper, we present our study on a total variation based gradient descent (TV-GD) algorithm for sparse-view PAI reconstruction. This algorithm involves the total variation (TV) method in compressed sensing (CS) theory. The objective function of the algorithm is modified by adding the TV value of the reconstructed image. With this modification, the reconstructed image could be closer to the real optical energy distribution map. Additionally in the proposed algorithm, the photoacoustic data is processed and the image is updated individually at each detection point. In this way, the calculation with large matrix can be avoided and a more frequent image update can be obtained. Through the numerical simulations, the proposed algorithm is verified and compared with other reconstruction algorithms which have been widely used in PAI. The peak signal-to-noise ratio (PSNR) of the image reconstructed by this algorithm is higher than those by the other algorithms. Additionally, the convergence of the algorithm, the robustness to noise and the tunable parameter are further discussed. The TV-based algorithm is also implemented in the in vitro experiment. The better performance of the proposed method is revealed in the experiments results. From the results, it is seen that the TV-GD algorithm may be a practical and efficient algorithm for sparse-view PAI reconstruction.     

A comparison of models used as alternative magnetic resonance image reconstruction methods

In magnetic resonance (MR) imaging, excellent reconstructions are obtained on large data sets using the inverse discrete Fourier transform (IDFT). Modeling procedures have been proposed to overcome the image artifacts from the truncation of small data sets. In this paper, a relationship between image reconstruction using modeling and the standard IDFT is presented. A comparison of the assumptions behind the Smith and Haacke models is given and an experimental evaluation of the validity of the models provided. Various methods of evaluating model coefficients are discussed. Images are reconstructed from both models using the Transient Error reconstruction approach (TERA) algorithm. TERA is an algorithm that reintroduces data information components that cannot be modeled: useful when the assumed model characteristics do not completely match all portions of the image. Although very different in their basic assumptions, both the Smith and Haacke models were found to reduce truncation artifacts and improve resolution when used with the TERA algorithm.     

三种降采样成像策略的性能优化以及与传统傅里叶望远镜的比较

为了获得成像质量较好且成像时间较少的新型傅里叶望远镜成像策略,本文比较了三种降采样成像策略(压缩感知方法 (CS)、低频全采样方法 (LF)和变密度随机采样方法 (VD))与传统傅里叶望远镜(FT)在图像质量和成像时间上的差异.分析方法如下:利用传统FT外场实验所获得的目标频谱数据作为基础,三种降采样方法 (LF, VD和CS)分别按照各自的采样模式和重构方法实现目标图像的重构;通过直观观察和Strehl比两种方法比较三种降采样方法与传统FT在图像质量上的差异;通过分析成像时间的组成要素,初步比较三种降采样方法与传统FT在成像时间上的差异.分析表明:1)压缩感知方法的图像质量优于其他两种降采样方法 (LF和VD),但略低于传统成像结果; 2)压缩感知方法在成像质量上略低于传统FT,但在成像时间上却明显小于传统FT; 3)分析中采用的外场数据均含噪声,这说明上述三种降采样重构过程对噪声有较好的鲁棒性.综合上述分析结果可以看出,基于压缩感知的傅里叶望远镜(CS-FT)是在实际含噪情况下可大幅减少成像时间的优良成像策略.