Future trends in 3D medical imaging

Areas where significant progress is needed are outlined. A unified conceptual framework based on generic tasks, showing the intricacies and the dependences that exist among completely distinct and intensive research lines, is defined. True 3-D volume imaging devices are then discussed, with the main emphasis on vascular network reconstruction. Segmentation issues are briefly reviewed. The fusion of information dealing with image, signal, and model is described. Simulation and planning problems are discussed.     

Phonon imaging in 3D with a fibre probe

We show for the first time that a single ultrasonic imaging fibre is capable of simultaneously accessing 3D spatial information and mechanical properties from m...     

Dual-contrast single breath-hold 3D abdominal MR imaging

Object: Multiple contrasts are often helpful for a comprehensive diagnosis. In 3D abdominal MRI, breath-hold techniques are preferred for single contrast acquisitions to avoid respiratory artifacts. In this paper, highly accelerated parallel MRI is used to acquire large 3D abdominal volumes with two different contrasts within a single breath-hold. Material and methods: In vivo studies have been performed on six healthy volunteers, combining T ₁- and T ₂-weighted, gradient- or spin-echo based scans, as well as water/fat resolved imaging in a single breath-hold. These 3D scans were acquired with an acceleration factor of six, using a prototype 32-element receive array. Results: The presented approach was tested successfully on all volunteers. The whole liver area was covered by a FOV of 350 × 250 × 200 mm³ for all scans with reasonable spatial resolution. Arbitrary scan protocols generating different contrasts have been shown to be combinable in this single breath-hold approach. Good spatial correspondence with negligible spatial offset was achieved for all different scan combinations acquired in overall breath-hold times between 15 and 25 s. Conclusion: Enabled by highly parallel imaging technology, this study demonstrates the technical feasibility and the promising image quality of single breath-hold dual contrast MRI.     

A survey of 3D medical imaging technologies

Three-dimensional medical imaging methodologies are surveyed with respect to hardware versus software, stand-alone versus on-the-scanner, speed, interaction, rendering methodology, fidelity, ease of use, cost, and quantitative capability. The question of volume versus surface rendering is considered in more detail. Research results are cited to illustrate the capabilities discussed.     

3D diffusion tensor imaging with 2D navigated turbo spin echo

A Cartesian two-dimensional navigator with variable orientation for online motion correction is introduced. It corrects for all possible zeroth- and first-order phase errors due to rigid-body motion of a subject during the diffusion-weighting preparation. The technique is developed for the application of three-dimensional (3D) imaging sequences, which offer the opportunity of high-resolution diffusion-weighted imaging, or diffusion tensor imaging (DTI) with isotropic voxel resolution. The navigator was applied to a displaced 3D turbo spin-echo sequence with an ECG-gated diffusion preparation to avoid phase errors due to gross brain pulsation. Online and offline corrected in vivo images acquired with this sequence are compared to investigate the advantages of online correction. Also eigenvector maps of the diffusion tensor are presented with an isotropic resolution of 1 mm³, which indicate that this new navigator technique is a promising approach for high-resolution DTI.     

Towards statistically optimal interpolation for 3D medical imaging

The use of a statistical estimation technique called kriging, which produces estimation error measurements and analyzes the volumetric grid to determine sample value variability, is described. The use of interpolation in 3D medical imaging are first reviewed. Several different interpolation techniques, including linear trilinear, and tricubic interpolation techniques, are described and assessed. The kriging statistical estimation process is presented, and the results of applying it to slice interpolation and surface visualization are reported. The results indicate the potential of kriging for interpolation in 3D medical imaging and point out the need for further work     

基于相关型图像传感器的3D AAMs人脸特征自动定位

在基于相关型图像传感器（corre lation image sensor,CIS）三维人脸成像的基础上,利用由CIS得到的深度信息和与之对应的亮度信息,提出一种建立三维人脸模型的方法,将二维AAMs扩展为三维AAMs,融合人脸的形状,纹理和深度信息来构建三维人脸模型。同时,采用三维曲率对人脸特征进行初步定位来解决AAMs初始姿态参数选择问题。人脸面部定位实验证明此方法在不同人脸姿态,表情和光照条件下定位效果要优于传统的2DAAMs。     

3D imaging of lightning channel and leader progression velocity

Three‐dimensional imaging of lightning channels by means of UHF interferometers is performed. To realize this, a field campaign was conducted with two observation sites in Darwin, Australia. The distance between the two sites is about 14 km, and the triangulation scheme is applied to estimate locations of UHF radiation sources. Since the UHF radiation is more likely during the leader progression than in the return stroke phase, imaging of lightning channels is obtained mainly for downward leaders. Velocities of downward leaders are estimated and statistics are presented for first and subsequent strokes. The result is consistent with those of former investigations obtained by optical observations. The statistics for the velocity of leader propagation inside thunderclouds are presented herein. Using optical measurement, velocity estimation of leader progressions inside thunderclouds is impossible, and the presented results are concluded to be the first results of their kind. © 2001 Scripta Technica, Electr Eng Jpn, 137(3): 22–28, 2001     

Navigator motion correction of diffusion weighted 3D SSFP imaging

Diffusion weighted (DW) 3D steady state MR (SSFP) head imaging technique using navigator echo’s motion correction is presented. This new scheme enables acquisition of DW images even at regions where severe susceptibility is present. Another advantage is the moderate gradient performance requirements. DW imaging methods are sensitive to any kind of motion, thus, most of these methods might suffer from bulk motion artifacts. The common solution to avoid motion artifacts in a 2D DW SSFP acquisition is multi averaging. To avoid the time consuming multi averaging, the new scheme, described here, utilizes navigator echo’s motion correction to remove respiratory bulk motion artifacts. At some brain regions, where the motion is governed by blood or CSF pulsation, the navigator motion correction fails. At these regions the correction is an interpolation of corrections from regions where the motion is particularly of the respiratory type. The combination of a 3D sequence with a navigator echo motion correction, enables acquisition of 10 DW slices within a time of 0:50-2:30 min.     

基于三维人脸成像系统的复数域人脸识别方法

三维人脸识别是模式识别和人工智能领域的研究热点。提出了一种利用人脸图像的二维灰度信息和三维深度信息进行人脸识别的算法。首先利用相关型图像传感器构成三维实时人脸成像系统,并将获得的三维人脸物理数据用复数形式表达。再将特征脸（Eigenface）方法和Fisherface方法拓展到复数域,提出了复数域特征脸方法和复数域Fisherface方法。改进的复数域人脸识别方法在三维实时成像系统的人脸数据库中进行的人脸识别实验表明,复数域的三维人脸识别方法明显优于传统的人脸识别方法。     

基于线阵列超声相控阵三维成像的实现研究

为进一步增强常规超声相控阵的检测能力,弥补二维检测图像对缺陷空间信息反映的单一性,获取更利于分析和评价的工业超声检测三维图像。采用携带编码器的一维线阵列换能器,借助Phascan平台及其二次开发模块采集原始扫查数据,合成带有位置信息的B扫图像,通过基于区域增长技术的混合绘制法实现缺陷的三维重建。人工缺陷对比试块的三维成像测量结果误差多小于5%,实验研究表明,该成像方法和系统能在空间上对缺陷形成一个直观体现,同时能够精确的反映缺陷的位置、形态以及尺寸等信息,为后续实验研究的开展奠定基础。     

Single-Pair FÖrster Resonance Energy Transfer With Multiparametric Excitation and Detection

Forster resonance energy transfer (FRET) between a donor and an acceptor dye molecule is a common method to study the distances between single molecules in living cells in the nanometer range. Quantitative distance measurements are difficult to obtain in spite of the strong distance dependency of the energy transfer efficiency. One problem is the incomplete fluorescent labeling of the molecules, which leads to the so-called zero-efficiency peak caused by FRET pairs with missing or nonfluorescing acceptor dyes. Other problems occur due to spectral bleed through, direct acceptor excitation, and the difficulty to obtain the quantum efficiencies of the dyes and the detection efficiencies of the corresponding detectors. In order to correct these defects, the donor as well as the acceptor are excited alternatingly using pulsed interleaved excitation. Time-correlated single-photon counting enables the measurement of fluorescence lifetime; hence, the FRET efficiency can also be derived from the decrease of the donor fluorescence lifetime. The universal time-tagged time-resolved data format allows to retrieve all the necessary information from the fluorescence photons detected during the measurement.     

Accelerating cardiac diffusion tensor imaging combining local low-rank and 3D TV constraint

Magnetic Resonance Materials in Physics, Biology and Medicine - Diffusion tensor magnetic resonance imaging (DT-MRI, or DTI) is a promising technique for invasively probing biological tissue...     

Mirror-enhanced scanning light-field microscopy for long-term high-speed 3D imaging with isotropic resolution

Various biological behaviors can only be observed in 3D at high speed over the long term with low phototoxicity. Light-field microscopy (LFM) provides an elegan...     

基于压缩感知的稀疏阵列近景微波三维成像

近景微波成像技术在隐藏武器检测、生物医学成像、无损检测等领域具有广阔的应用前景.将压缩感知理论应用于近景微波成像系统,提出了基于稀疏阵列天线的三维成像几何及相应的回波模型,该几何通过高度向的稀疏阵列电子扫描、方位向的机械扫描及距离向发射宽带信号来实现三维分辨;文中将压缩感知理论应用于稀疏阵列三维成像,提出了基于压缩感知...     

Real-time imaging and triggering of 3D contrast-enhanced MR angiograms using MR fluoroscopy

Although a variety of timing techniques and data acquisition strategies have been, used for three-dimensional contrast-enhanced MR angiography, many are still limited by inadequate overall reliability, limited spatial resolution, or complexity. A technique is presented in this work in which contrast arrival is detected in the targeted vasculature in real time using MR fluoroscopy. Upon detection the operator triggers a 3D MR angiographic acquisition which uses an elliptical centric view order. It is shown that the view order intrinsically provides a high degree of venous suppression which in turn allows acquisition times of 30 s or longer. permitting high spatial resolution. The reliability of fluoroscopic triggering in bolus detection is shown to be over 90%. The technique provides high quality contrast-enhanced MR angiograms for many vascular regions.