A maximum likelihood approach to parallel imaging with coil sensitivity noise

Parallel imaging is a powerful technique to speed up magnetic resonance (MR) image acquisition via multiple coils. Both the received signal of each coil and its sensitivity map, which describes its spatial response, are needed during reconstruction. Widely used schemes such as SENSE assume that sensitivity maps of the coils are noiseless while the only errors are in coil outputs. In practice, however, sensitivity maps are subject to a wide variety of errors. At first glance, sensitivity noise appears to result in an errors-in-variables problem of the kind that is typically solved using total least squares (TLSs). However, existing TLS algorithms are in general inappropriate for the specific type of block structure that arises in parallel imaging. In this paper, we take a maximum likelihood approach to the problem of parallel imaging in the presence of independent Gaussian sensitivity noise. This results in a quasi-quadratic objective function, which can be efficiently minimized. Experimental evidence suggests substantial gains over conventional SENSE, especially in nonideal imaging conditions like low signal-to-noise ratio (SNR), high g-factors and large acceleration, using sensitivity maps suffering from misalignment, ringing, and random noise.     

Efficient Spatially-Variant Single-Pixel Imaging Using Block-Based Compressed Sensing

Single-pixel imaging is an important alternative to conventional camera. Only a single-pixel detector is needed to capture image data by measuring the correlation of the target scene and a series of sensing patterns. Conventionally, Nyquist-Shannon theorem requires measurements not less than the image pixels for an error-free reconstruction. Compressed sensing (CS) enables image reconstructions with fewer measurements but the image quality and computational cost remain the primary concerns. This paper presents an efficient single-pixel imaging technique based on blocked-based CS in which the sensing matrices are designed based on spatially-variant resolution (SVR). The proposed method decreases the number of measurements as well as the image reconstruction time using the SVR sensing patterns. Furthermore, it takes advantage of block-based CS to reduce the expenses of computational resources. The proposed method is evaluated and compared to conventional uniform resolution (UR) image reconstruction in terms of image quality and reconstruction time. The results show that the proposed method consistently reduces the reconstruction time and able to give better image quality at lower sampling ratio (SR). This provides an efficient reconstruction for single-pixel imaging which is desirable in practical application and situations where low sampling rate is required.

     

基于深度学习的单像素相机超分辨率成像优化

基于压缩感知和单像素成像的基本原理,设计了一种用于图像超分辨率重建的新型深度卷积神经网络架构.这种单像素超分辨率成像算法成功地将深度学习图像超分辨率重建技术与压缩感知单像素成像技术相结合,从而发展出一种全新的深度学习单像素成像优化方法.与传统的常规压缩感知图像重构算法相比,该算法有效提升了图像超分辨率重建精度和单像素成像质量.通过图像重建的仿真实验和单像素相机的成像实验验证,结果表明这种基于深度学习的新型单像素相机成像方式具有良好的性能表现.     

Real-Time Reconstruction of Sensitivity Encoded Radial Magnetic Resonance Imaging Using a Graphics Processing Unit

A barrier to the adoption of non-Cartesian parallel magnetic resonance imaging for real-time applications has been the times required for the image reconstructions. These times have exceeded the underlying acquisition time thus preventing real-time display of the acquired images. We present a reconstruction algorithm for commodity graphics hardware (GPUs) to enable real time reconstruction of sensitivity encoded radial imaging (radial SENSE). We demonstrate that a radial profile order based on the golden ratio facilitates reconstruction from an arbitrary number of profiles. This allows the temporal resolution to be adjusted on the fly. A user adaptable regularization term is also included and, particularly for highly undersampled data, used to interactively improve the reconstruction quality. Each reconstruction is fully self-contained from the profile stream, i.e., the required coil sensitivity profiles, sampling density compensation weights, regularization terms, and noise estimates are computed in real-time from the acquisition data itself. The reconstruction implementation is verified using a steady state free precession (SSFP) pulse sequence and quantitatively evaluated. Three applications are demonstrated; real-time imaging with real-time SENSE 1) or $k$-$t$ SENSE 2) reconstructions, and 3) offline reconstruction with interactive adjustment of reconstruction settings.      

基于三维直方图的Fisher评价函数图像分割方法

提出了一种基于三维直方图的Fisher评价函数图像分割方法,该方法利用了图像像素点的灰度信息和邻域的均值信息及中值信息,给出了三维阈值分割方法。实验结果表明,相比较基于二维直方图的Fisher线性判别函数图像分割方法,在图像受高斯噪声干扰的情况下,该方法具有更好的分割效果。     

Attenuation compensation of cone beam SPECT images using maximum likelihood reconstruction

Attenuation compensation for cone beam single-photon emission computed tomography (SPECT) imaging is performed by cone beam maximum likelihood reconstruction with attenuation included in the transition matrix. Since the transition matrix is too large to be stored in conventional computers, the E-M maximum likelihood estimator is implemented with a ray-tracing algorithm, efficiently recalculating each matrix element as needed. The method was applied and tested in both uniform and nonuniform density phantoms. Test projections sets were obtained from Monte Carlo simulations and experiments using a commercially available cone beam collimator. For representative regions of interest. reconstruction of a uniform sphere is accurate to within 3% throughout, in comparison to a reference image simulated and reconstructed without attenuation. High- and low-activity regions in a uniform density are reconstructed accurately, except that low-activity regions in a more active background have a small error. This error is explainable by the nonnegativity constraints of the E-M estimator and the image statistical noise     

基于多方向小波提升IRFPA盲元检测精度方法

红外焦平面成像质量受材料生长及器件制备工艺的影响,易出现盲元、条纹噪声等缺陷。条纹噪声经常会导致盲元的检测偏差,准确的盲元检测对于后续图像处理具有重要意义。利用双密度双树复数小波分解的多方向性小波系数,结合广义高斯分布将高频小波系数按照对条纹噪声影响程度分别赋予不同权值并进行单支重构,消除了条纹噪声对盲元检测的影响,得到初步干净的预处理图像,进而对预处理图像运用3准则进行盲元检测。通过短波HgCdTe红外焦平面成像的实践验证,该方法对具有条纹噪声特征的红外图像盲元检测更加准确。     

一种实用的小目标配准方法

图像配准是图像融合技术的基本环节和首要问题,只有经过配准后的图像才能进行有效的融合。其中,小目标由于几乎无特征信息可以利用,所以常规的配准方法都不适用。针对图像识别中小目标的配准问题,分析了其配准特点,创新性地提出了先配准目标视场,再配准目标位置的方法,提出了视场配准的概念。首先运用成像原理,用焦距、分辨率和像元尺寸建立不同CCD之间的视场对应关系,利用此关系完成目标视场的截取放大,使不同CCD得到的图像视场一样大。然后在分析通常采用的最小平均绝对误差（MAD）相关匹配方法缺陷的基础上,提出用最多近邻点距离（MCD）的匹配方法来对准目标位置,完成目标质心的配准。实验结果表明,此方法可以很好地配准小目标,且误差不超过2个像素。由于其针对性强．因而具有较强的实际应用价值。     

Multi-channel microstrip transceiver arrays using harmonics for high field MR imaging in humans

Radio-frequency (RF) transceiver array design using primary and higher order harmonics for in vivo parallel magnetic resonance imaging imaging (MRI) and spectroscopic imaging is proposed. The improved electromagnetic decoupling performance, unique magnetic field distributions and high-frequency operation capabilities of higher-order harmonics of resonators would benefit transceiver arrays for parallel MRI, especially for ultrahigh field parallel MRI. To demonstrate this technique, microstrip transceiver arrays using first and second harmonic resonators were developed for human head parallel imaging at 7T. Phantom and human head images were acquired and evaluated using the GRAPPA reconstruction algorithm. The higher-order harmonic transceiver array design technique was also assessed numerically using FDTD simulation. Compared with regular primary-resonance transceiver designs, the proposed higher-order harmonic technique provided an improved g-factor and increased decoupling among resonant elements without using dedicated decoupling circuits, which would potentially lead to a better parallel imaging performance and ultimately faster and higher quality imaging. The proposed technique is particularly suitable for densely spaced transceiver array design where the increased mutual inductance among the elements becomes problematic. In addition, it also provides a simple approach to readily upgrade the channels of a conventional primary resonator microstrip array to a larger number for faster imaging.     

噪声密度不敏感的随机采样椒盐噪声滤波算法

中值滤波系列算法在处理被不同密度椒盐噪声污染的细节图像和平坦图像时,降噪性能不一致。本文借鉴开关中值滤波和压缩感知的思想,提出了随机采样滤波算法去除椒盐噪声。算法以噪声检测为基础,将被椒盐噪声污染的图像分为疑似噪声像素和信号像素,随机采样仅对信号像素采样。然后,利用正交匹配追踪算法重构出被污染前的图像,替代了中值滤波对噪声像素的估计。由于随机采样滤波基于压缩感知理论,对稀疏信号的重构具有最少测量次数的条件,因此随机采样点的数量具有一定的浮动空间,表现为对噪声密度不敏感。以被不同噪声密度污染图像的纹理、平坦局部区域进行验证,实验表明,当噪声密度在一定范围内变化时,算法可以实现对噪声密度不敏感。在高密度噪声污染的情况下,相较于中值滤波系列算法,随机采样滤波算法具有更好的细节保留能力和滤波能力。对标准测试图像进行了全局滤波,不同噪声密度具有一致的滤波效果,与自适应滤波算法相比,随机采样滤波算法在处理包含密集边缘特征的区域时更具备优势。      

基于DBSCAN的图像分割算法研究

图像分割能将一幅数字图像分成多个的不同区域,是计算机视觉的主要研究领域之一。文中在系统性研究DBSCAN算法的基础上，提出了一种改进型DBSCAN图像分割方法。该方法首先计算图像中每个像素点的局部密度，然后通过寻找局部密度峰值点来确定核心点，同时将邻域内的像素点加入同一簇中，来处理不同密度区域和噪声点的影响。实验结果表明，该算法对参数敏感度较低，能有效处理不同密度区域和噪声点，相比标准DBSCAN图像分割方法，其在聚类正确率、精度和效率等方面的表现更优秀。     

基于辐射特征的干扰弹检测方法研究

根据干扰弹在起燃时间内辐射能量急剧增加的特点,利用图像序列中图像灰度斜率的突变进行干扰检测是一种比较有效的方法。利用仿真选择有效的数据平滑方法对灰度变化斜率进行估计,以消除焦平面探测器采样特性和噪声的影响,定量分析了影响干扰弹检测概率的各种因素,评价了利用辐射特征检测干扰弹的优点与不足。     

基于稀疏和冗余表象的鬼成像雷达研究进展

基于稀疏和冗余表象的鬼成像雷达(Ghost Image via Sparsity Constraints,GISC Lidar)是一种结合光场空间涨落特性和现代信息论的全新雷达成像体制,其成像视场和分辨率无关,由此可在探测时采用大视场凝视成像模式捕捉运动目标以对其进行高分辨率成像探测。与闪光照相雷达需要将目标的反射光信号成像分布在焦平面阵列光电探测器件上相比,GISC雷达只需要一个无空间分辨能力的单像素探测器接收目标场景的全部反射光信号,因此可以极大地提升系统的成像探测灵敏度。此外,GISC雷达在成像探测过程中可以利用图像的各种先验约束,从而突破奈奎斯特采样定理对采样次数的要求,大幅度提高图像的信息获取效率。文中将结合上海光机所将鬼成像技术应用于雷达探测的研究历程,介绍GISC雷达研究进展,并指出GISC雷达工程化实际应用中仍待解决的若干问题。     

基于精确模型和逆投影的超大视场红外图像畸变校正

针对超大视场红外图像畸变大、与人眼视觉差异明显的问题,提出了一种基于精确模型和逆投影的超大视场红外图像畸变校正算法以改善其视觉效果.该算法首先利用精确模型对超大视场红外相机成像中的物、像关系进行描述;然后,针对红外图像像素采样率不高的缺点,利用较为精确的三次卷积插值法对图像进行插值来补全成像信息;最后,根据校正图像上的待赋值像点的坐标,结合校正模型和超大视场红外相机精确模型,计算该像点逆投影到插值图像时的对应坐标,并以最近邻像点像素值作为校正后图像像点的赋值.车载道路场景下的超大视场红外图像畸变校正实验结果显示,所提出的算法图像校正结果边界清晰、无锯齿效应,对场景中的直线平均还原偏差小于0.35 pixels,表明该算法对超大视场红外图像畸变校正具有较好的适用性.     

Predictive (un)distortion model and 3-D reconstruction by biplane snakes

This paper is concerned with the three-dimensional (3-D) reconstruction of coronary vessel centerlines and with how distortion of X-ray angiographic images affects it. Angiographies suffer from pincushion and other geometrical distortions, caused by the peripheral concavity of the image intensifier (II) and the nonlinearity of electronic acquisition devices. In routine clinical practice, where a field-of-view (FOV) of 17-23 cm is commonly used for the acquisition of coronary vessels, this distortion introduces a positional error of up to 7 pixels for an image matrix size of 512 x 512 and an FOV of 17 cm. This error increases with the size of the FOV. Geometrical distortions have a significant effect on the validity of the 3-D reconstruction of vessels from these images. We show how this effect can be reduced by integrating a predictive model of (un)distortion into the biplane snakes formulation for 3-D reconstruction. First, we prove that the distortion can be accurately modeled using a polynomial for each view. Also, we show that the estimated polynomial is independent of focal length, but not of changes in anatomical angles, as the II is influenced by the earth's magnetic field. Thus, we decompose the polynomial into two components: the steady and the orientation-dependent component. We determine the optimal polynomial degree for each component, which is empirically determined to be five for the steady component and three for the orientation-dependent component. This fact simplifies the prediction of the orientation-dependent polynomial, since the number of polynomial coefficients to be predicted is lower. The integration of this model into the biplane snakes formulation enables us to avoid image unwarping, which deteriorates image quality and therefore complicates vessel centerline feature extraction. Moreover, we improve the biplane snake behavior when dealing with wavy vessels, by means of using generalized gradient vector flow. Our experiments show that the proposed methods in this paper decrease up to 88% the reconstruction error obtained when geometrical distortion effects are ignored. Tests on imaged phantoms and real cardiac images are presented as well.     

Computational Reconstruction of Three-Dimensional Integral Imaging by Rearrangement of Elemental Image Pixels

In this paper, we present a method to implement computational three-dimensional (3D) integral imaging (II). This method is based on Pixels of the Elemental Image Rearrangement Technique (PERT). In our proposed method for computational reconstruction of II, the reconstructed 3D image is obtained by using the entire elemental images which are captured from the lenslet array. Instead of averaging the elemental images, our proposed method rearranges pixels of each elemental image. Therefore, the reconstructed 3D image has the same number of pixels as the entire elemental images' pixels. To verify this computational reconstruction method, we have implemented optical experiments.      

代数迭代法在激光反射断层成像目标重构中的应用

为了提高激光反射断层成像目标重构的图像质量,在目前激光反射断层成像普遍采用反投影算法重构图像的基础上,将CT成像中常用的迭代重建算法引入到激光反射断层成像的图像重构过程中。分析了反投影算法中的直接反投影、R-L和S-L滤波反投影以及迭代重建算法在图像重构中的性能特性。进行了仿真和外场实验,结果表明:在直接反投影基础上添加了滤波器的反投影算法在减小误差和抑噪能力上都明显提高;另外相比于反投影算法,代数迭代重建算法表现出更好的重建质量,且具有更强的抑噪性能。     

Expectation maximization reconstruction of positron emissiontomography images using anatomical magnetic resonance information

Using statistical methods the reconstruction of positron emission tomography (PET) images can be improved by high-resolution anatomical information obtained from magnetic resonance (MR) images. The authors implemented two approaches that utilize MR data for PET reconstruction. The anatomical MR information is modeled as a priori distribution of the PET image and combined with the distribution of the measured PET data to generate the a posteriori function from which the expectation maximization (EM)-type algorithm with a maximum a posteriori (MAP) estimator is derived. One algorithm (Markov-GEM) uses a Gibbs function to model interactions between neighboring pixels within the anatomical regions. The other (Gauss-EM) applies a Gauss function with the same mean for all pixels in a given anatomical region. A basic assumption of these methods is that the radioactivity is homogeneously distributed inside anatomical regions. Simulated and phantom data are investigated under the following aspects: count density, object size, missing anatomical information, and misregistration of the anatomical information. Compared with the maximum likelihood-expectation maximization (ML-EM) algorithm the results of both algorithms show a large reduction of noise with a better delineation of borders. Of the two algorithms tested, the Gauss-EM method is superior in noise reduction (up to 50%). Regarding incorrect a priori information the Gauss-EM algorithm is very sensitive, whereas the Markov-GEM algorithm proved to be stable with a small change of recovery coefficients between 0.5 and 3%     

基于Zernike-Facet模型和总体最小二乘的弱小目标检测

弱小目标一般是图像局部区域的极值点。针对这个特点,依据二元三次函数的极值理论,该文提出了一种新的弱小目标候选点的检测方法。发展了一种新的图像局部灰度拟合模型,即Zernike-facet模型,模型参数的求解采用比最小二乘(LS)抗噪能力更强的总体最小二乘(TLS)算法。新检测方法通过Zernike-facet模型和TLS对原始图像中每一个像素的局部区域进行曲面拟合,然后在拟合曲面上提取极值点作为目标候选点。仿真表明,新方法在抑制噪声上优于其他常用方法。可见光/红外图像小目标检测实验也证实了新方法的有效性。     

A Natural Pixel Decomposition for Two-Dimensional Image Reconstruction

In two-dimensional image reconstruction from line integrals using maximum likelihood, Bayesian, or minimum variance algorithms, the x-y plane on which the object estimate is defined is decomposed into nonoverlapping regions, or "pixels." This decomposition of an otherwise continuous structure results in significant errors, or model noise, which can exceed the effects of the fundamental measurement noise.