基于传感器参数和目标轮廓中心的自动配准算法研究

通过对光电成像型反舰导弹的成像过程分析,提出了一种自动配准算法。其基本思路是将图像变换模型分解,逐步简化。可见光和红外图像配准时的变换为仿射变换。首先,利用传感器参数的调整,消除图像间的比例变化,将仿射变换简化为刚体变换;然后,依赖图像信息,用形态学边缘检测的方法求取目标的轮廓中心,以此为控制点消除图像间的平移变化,实现图像的完全配准;最后,通过观察海天线是否重合以及利用均方根误差原则,对算法的配准效果做详细评估。仿真实验表明,该算法准确、快速,配准精度满足目标识别的要求,可以较好地解决异类传感器弱小目标图像配准的难题。     

Multicolour analysis and local image correlation in confocal microscopy

Multiparameter fluorescence microscopy is often used to identify cell types and subcellular organelles according to their differential labelling. For thick objects, the quantitative comparison of different multiply labelled specimens requires the three-dimensional (3-D) sampling capacity of confocal laser scanning microscopy, which can be used to generate pseudocolour images. To analyse such 3-D data sets, we have created pixel fluorogram representations, which are estimates of the joint probability densities linking multiple fluorescence distributions. Such pixel fluorograms also provide a powerful means of analysing image acquisition noise, fluorescence cross-talk, fluorescence photobleaching and cell movements. To identify true fluorescence co-localization, we have developed a novel approach based on local image correlation maps. These maps discriminate the coincident fluorescence distributions from the superimposition of noncorrelated fluorescence profiles on a local basis, by correcting for contrast and local variations in background intensity in each fluorescence channel. We believe that the pixel fluorograms are best suited to the quality control of multifluorescence image acquisition. The local image correlation methods are more appropriate for identifying co-localized structures at the cellular or subcellular level. The thresholding of these correlation maps can further be used to recognize and classify biological structures according to multifluorescence attributes.     

Image automatic mosaics based on contour phase correlation

The image planar mosaics is studied, and an image automatic mosaics algorithm on the basis of contour phase correlation is proposed in this paper. To begin with, by taking into account mere translations and rotations between images, a contour phase correlation algorithm is used to realize the preliminary alignments of images, and the initial projective transformation matrices are obtained. Then, an optimization algorithm is used to optimize the initial projective transformation matrices, and complete the precise image mosaics. The contour phase correlation is an improvement on the conventional phase correlation in two aspects: First, the contours of images are extracted, and the phase correlation is applied to the contours of images instead of the whole original images; Second, when there are multiple peak values approximate to the maximum peak value in the δ function array, their corresponding translations can be regarded as candidate translations and calculated separately, and the best translation can be determined by the optimization of conformability of two images in the overlapping area. The running results show that the proposed algorithm can consistently yield high-quality mosaics, even in the cases of poor or differential lighting conditions, existences of minor rotations, and other complicated displacements between images. Translated from Journal of Dalian University of Technology, 2005, 45(1): 68–74 [译自: 大连理工大学学报]     

Differential interference contrast and confocal reflectance imaging of collagen organization in three-dimensional matrices

The remodeling of extracellular matrices by cells plays a defining role in developmental morphogenesis and wound healing as well as in tissue engineering. Three-dimensional (3-D) type I collagen matrices have been used extensively as an in vitro model for studying cell-induced matrix reorganization at the macroscopic level. However, few studies have directly assessed the process of 3-D extracellular matrix (ECM) remodeling at the cellular and subcellular level. In this study, we directly compare two imaging modalities for both quantitative and qualitative imaging of 3-D collagen organization in vitro: differential interference contrast (DIC) and confocal reflectance imaging. The results demonstrate that two-dimensional (2-D) DIC images allow visualization of the same population of collagen fibrils as observed in 2-D confocal reflectance images. Thus, DIC can be used for qualitative assessment of fibril organization, as well as tracking of fibril movement in sequential time-lapse 2-D images. However, we also found that quantitative techniques that can be applied to confocal reflectance images, such as Fourier transform analysis, give different results when applied to DIC images. Furthermore, common techniques used for 3-D visualization and reconstruction of confocal reflectance datasets are not generally applicable to DIC. Overall, obtaining a complete understanding of cell-matrix mechanical interactions will likely require a combination of both wide-field DIC imaging to study rapid changes in ECM deformation which can occur within minutes, and confocal reflectance imaging to assess more gradual changes in cell-induced compaction and alignment of ECM which occur over a longer time course.     

Precise 3D image alignment in micro-axial tomography

Micro (µ‐) axial tomography is a challenging technique in microscopy which improves quantitative imaging especially in cytogenetic applications by means of defined sample rotation under the microscope objective. The advantage of µ‐axial tomography is an effective improvement of the precision of distance measurements between point‐like objects. Under certain circumstances, the effective (3D) resolution can be improved by optimized acquisition depending on subsequent, multi‐perspective image recording of the same objects followed by reconstruction methods. This requires, however, a very precise alignment of the tilted views. We present a novel feature‐based image alignment method with a precision better than the full width at half maximum of the point spread function. The features are the positions (centres of gravity) of all fluorescent objects observed in the images (e.g. cell nuclei, fluorescent signals inside cell nuclei, fluorescent beads, etc.). Thus, real alignment precision depends on the localization precision of these objects. The method automatically determines the corresponding objects in subsequently tilted perspectives using a weighted bipartite graph. The optimum transformation function is computed in a least squares manner based on the coordinates of the centres of gravity of the matched objects. The theoretically feasible precision of the method was calculated using computer‐generated data and confirmed by tests on real image series obtained from data sets of 200 nm fluorescent nano‐particles. The advantages of the proposed algorithm are its speed and accuracy, which means that if enough objects are included, the real alignment precision is better than the axial localization precision of a single object. The alignment precision can be assessed directly from the algorithm's output. Thus, the method can be applied not only for image alignment and object matching in tilted view series in order to reconstruct (3D) images, but also to validate the experimental performance (e.g. mechanical precision of the tilting). In practice, the key application of the method is an improvement of the effective spatial (3D) resolution, because the well‐known spatial anisotropy in light microscopy can be overcome. This allows more precise distance measurements between point‐like objects.     

基于Otsu准则和直线截距直方图的阈值分割

对二维Otsu法中类间离散度测度进行了分析,发现按该算法对被噪声污染图像的二维直方图进行划分时,所得两类的类内均值点容易远离主对角线,因而抗噪声能力不足。针对以上情况,本文提出了一种新算法,该算法基于二维直方图中直线阈值分割的思想,利用像素点的二维信息直接建立阈值直线的截距直方图;然后应用Otsu准则对该一维直方图求解最佳截距阈值,并应用该阈值和二维信息完成图像分割。对提出的算法与传统二维Otsu法进行了比较和分析,结果表明:提出的算法可以有效避免传统算法在抗噪方面的缺陷,当实验图像的噪声方差大于0.003且逐渐增加时,提出的算法抗噪表现稳健;另外,提出的算法计算阈值的速度比基于二维Otsu法的直分法和直线阈值法快2个数量级以上,占用内存空间更少。因而提出的算法是一种抗噪稳健且快速有效的阈值分割算法,更适于实时应用。     

基于视差与基线距相关三目立体匹配法

采用统计过程控制方法对二维图像特征点区域定位，并提取二维图像特征点。避免了提取二维图像特征点时，根据被处理图像的先验信息，利用试探方法确定阈值的局限性。在立体匹配时，将灰度相关系数小于最大灰度相关系数一定范围内的特征点作为灰度相关复峰初始匹配特征点集合。根据由正确匹配特征点组成的视差矩阵与对应的基线距矩阵存在极大相关性，从灰度相关复峰初始匹配特征点集合中确定唯一匹配特征点。通过对外形复杂的实际物体及已知精确三维坐标的标准工件的三维重建，证实了文中所提方法的有效性和可靠性。     

Scanning electron microscope charging effect model for chromium/quartz photolithography masks

We propose a new method for fitting a model of specimen charging to scanning electron microscope (SEM) images. Charging effects cause errors when one attempts to infer the size or shape of a specimen from an image. The goal of our method is to enable image analysis algorithms for measurement, segmentation, and three-dimensional (3-D) reconstruction that would otherwise fail on images containing charging effects. Our model is applied to images of chromium/quartz photolithography masks and may also work in the more general case of isolated metal islands on a flat insulating substrate. Unlike methods based on Monte Carlo simulation, our simulation method does not handle more general topographies or specimens composed entirely of an insulator; it is a crude approximation to the physical charging process described in more detail in Cazaux (1986) and Melchinger and Hofmann (1985), but can be fit with quantitative accuracy to real SEM images. We only consider changes in intensity and do not model charging-induced distortion of image coordinates. Our approach has the advantage over existing methods of enabling fast prediction of charging effects so it may be more practical for image analysis applications.     

Confocal microscopic image sequence compression using vector quantization and three-dimensional pyramid

The three-dimensional (3-D) pyramid compressor project at the University of Glasgow has developed a compressor for images obtained from confocal laser scanning microscopy (CLSM) device. The proposed method using a combination of image pyramid coder and vector quantization techniques has good performance at compressing confocal volume image data. An experiment was conducted on several kinds of CLSM data using the presented compressor compared with other well-known volume data compressors, such as MPEG-1. The results showed that the 3-D pyramid compressor gave a higher subjective and objective image quality of reconstructed images at the same compression ratio, and presented more acceptable results when applying image processing filters on reconstructed images.     

Application of confocal scanning laser microscopy for an automated nuclear grading of prostate lesions in three dimensions

Confocal scanning laser microscopy provides the opportunity to obtain three-dimensional (3-D) images by piling up consecutive confocal planes. This technique was applied to capture 3-D images from 100-μm-thick tissue blocks from prostate lesions (hyperplasia, dysplasia, adenocarcinomas). Automated methods were implemented to perform a nuclear grading of 3-D cell nuclei from these specimens. Special attention was focused on the development of a new approach to 3-D chromatin texture analysis. This method uses mathematical morphology operations to tessellate the chromatin into homogeneous domains. The nuclear features (volume, shape, texture) were subjected to a discriminant analysis. Using a set of five features, the classification of cell nuclei yielded an accuracy of 963%. The results indicate the potential of 3-D imaging and analysis techniques for an automated nuclear grading of prostate lesions.     

Improvement in the resolution of three-dimensional data sets collected using optical serial sectioning

In this paper an approach for improving the quality of 3-D microscopic images obtained through optical serial sectioning is described and implemented. A serially sectioned image is composed of a sequence of 2-D images obtained by incrementing the focusing plane of the microscope through the specimen of interest; ideally, the image obtained at each focusing plane should be in focus, and should contain information lying only within that plane. In practice, however, the images obtained contain redundant information from neighbouring focusing planes and are blurred by a three-dimensional low-pass distortion. These degradations are a consequence of the limited aperture of any optical system; using principles of geometric optics and allowing for the passage of light through the specimen, we are able to demonstrate that the microscope distortion can be described as a linear system, if the absorption of the specimen is assumed to be linear and non-diffractive. The transfer function of the microscope is found to zero a biconic region of 3-D spatial frequencies orientated along the optical axis; a closed-form expression is derived for the low-pass transfer function of the microscope outside the region of missing frequencies. The planar resolution of the serial sections can be greatly improved by convolving the image obtained with the inverse of the low-pass distortion function, although the missing cone of frequencies is not recoverable. The reconstruction technique is demonstrated using both simulated images, to demonstrate more clearly the effects of the distortion and the accuracy of the subsequent reconstruction, and actual experiments with a pollen grain and a stained preparation of human cerebellum tissue.     

Image analysis of particle dispersions in microscopy images of cryo-sectioned sausages

Two feature extraction methods, the three-dimensional (3-D) local box-counting method and the area distribution method, are presented to describe the fat dispersion pattern on digital microscopy images of cryo-sectioned sausages. Both methods calculate whole arrays of variables for each microscopy image. The 3-D box-counting method calculates scale dependent (local) dimensions. This is in contrast to common fractal methods, which are univariate. Principal component analysis (PCA) was used to show that different sausages yield different fat dispersion patterns. Partial least square regression (PLS) shows that there is a correlation between the variables gained with both methods and the fat content.     

A new algorithm to align three-dimensional maps of helical structures

By subtracting one three-dimensional (3-D) map from another, one can calculate a difference map that can reveal structural changes, such as conformational changes, not detectable by eye. Furthermore, statistical significances can be assigned to such differences. The validity of the features in the difference map, however, depends on the alignment of the two maps; that is, one needs to align the two 3-D maps so that densities corresponding to equivalent parts of the structures are at the same coordinates. An existing method using the Fourier-Bessel coefficients Gn,1(R) is commonly used for the alignment of maps of helical structures. This procedure works well if the two maps have most features in common. But if they do not, it is difficult to control which features are used in the alignment procedure since the contributions from different features in the map are not easy to separate. We devised a procedure using the radial transform of Gn,1(R) (i.e., g(n),1(r)), which retains the powerful mathematical advantage of the Fourier-Bessel representation of the data and which provides the ability to select the radial features used in the alignment procedure. We applied the new method to 3-D maps of F-actin and F-actin decorated with various myosin motor constructs. Whereas the procedure using G(n),1(R) failed to align myosin-S1 decorated actin to undecorated actin, the new procedure accurately aligned maps.     

Three-dimensional molecular distribution in single cells analysed using the digital imaging microscope

Cellular changes in molecular distribution are believed to underly a wide range of cell functions. In order to investigate changes in molecular distribution in single cells utilizing fluorescent probes we have developed a digital imaging microscope. The system, consisting of both hardware and software, automatically acquires 3-D data sets consisting of optical sections and then processes such data to facilitate the analysis of molecular distribution in single cells. The first major step in processing reverses distortion introduced principally by the optics of the fluorescent microscope. Various procedures for accomplishing this task are compared and a method based on regularization theory is shown to give superior results for several different 3-D images. Following this step features of interest are automatically extracted from 3-D images utilizing an artificial 3-D visual system. This artificial visual system utilizes a system of spatial filters to identify regional characteristics of images, the information obtained from these filters being used to identify and characterize clusters of molecules within the image. This information is then utilized to construct a 3-D graphical model of molecular distribution in single cells. Such models are displayed in 3-D and may be further analysed utilizing interactive 3-D computer graphics. These methods are illustrated by results obtained regarding alpha-actinin distribution in single smooth muscle cells.     

Robust, globally consistent and fully automatic multi-image registration and montage synthesis for 3-D multi-channel images

The need to map regions of brain tissue that are much wider than the field of view of the microscope arises frequently. One common approach is to collect a series of overlapping partial views, and align them to synthesize a montage covering the entire region of interest. We present a method that advances this approach in multiple ways. Our method (1) produces a globally consistent joint registration of an unorganized collection of three-dimensional (3-D) multi-channel images with or without stage micrometer data; (2) produces accurate registrations withstanding changes in scale, rotation, translation and shear by using a 3-D affine transformation model; (3) achieves complete automation, and does not require any parameter settings; (4) handles low and variable overlaps (5-15%) between adjacent images, minimizing the number of images required to cover a tissue region; (5) has the self-diagnostic ability to recognize registration failures instead of delivering incorrect results; (6) can handle a broad range of biological images by exploiting generic alignment cues from multiple fluorescence channels without requiring segmentation and (7) is computationally efficient enough to run on desktop computers regardless of the number of images. The algorithm was tested with several tissue samples of at least 50 image tiles, involving over 5000 image pairs. It correctly registered all image pairs with an overlap greater than 7%, correctly recognized all failures, and successfully joint-registered all images for all tissue samples studied. This algorithm is disseminated freely to the community as included with the Fluorescence Association Rules for Multi-Dimensional Insight toolkit for microscopy (http://www.farsight-toolkit.org).     

Absorption and scattering correction in fluorescence confocal microscopy

In three-dimensional (3-D) fluorescence images produced by a confocal scanning laser microscope (CSLM), the contribution of the deeper layers is attenuated due to absorption and scattering of both the excitation and the fluorescence light. Because of these effects a quantitative analysis of the images is not always possible without restoration. Both scattering and absorption are governed by an exponential decay law. Using only one (space-dependent) extinction coefficient, the total attenuation process can be described. Given the extinction coefficient we calculate within a non-uniform object the relative intensity of the excitation light at its deeper layers. We also give a method to estimate the extinction coefficients which are required to restore 3-D images. An implementation of such a restoration filter is discussed and an example of a successful restoration is given.     

Data-driven, simultaneous blur and image restoration in 3-D fluorescence microscopy

A novel algorithm for simultaneous blur and image restoration (SBIR)* in three-dimensional (3-D) fluorescence microscopy is presented. All the internal parameters including the point spread function essential for the restoration are estimated from the data. Validation of the SBIR algorithm using simulated signals/images and known real world specimens is provided. Both lateral and axial resolution of images are improved by the application of the algorithm. Finally, the results of the application of the algorithm to unknown specimens are shown, demonstrating the potential of the algorithm in practical applications. Furthermore, evidence is provided to show that this algorithm can provide a turn-key system to deblur images in 3-D fluorescence microscopy.     

Cryoultramicrotomy of muscle: improved preservation and resolution of muscle ultrastructure using negatively stained ultrathin cryosections

Ultrathin sections of rapidly frozen, briefly pre-treated muscle tissue are cut and thereafter are thawed and contrasted using a negative staining technique. The method has provided micrographs in which the in-vivo order in the muscle fibres has been preserved well enough to enable both a more complete interpretation of X-ray diffraction evidence from muscle, and also a gain of new ultrastructural information on aspects of myofibril and myofilament architecture in different types of fibre. Examples here are taken from chicken, rabbit and fish muscles and show both the M-band and the bridge region of the A-band in great detail. To enhance the detail in the original images, one-dimensional (1-D) and 2-D averaging techniques (lateral smearing and step averaging, respectively) are used. Although there is major shrinkage in section thickness to about one-third of its original value, demonstrated here for the first time is the fact that the characteristic A-band lattice planes are preserved in these sections in 3-D. This confirms the usefulness of cryosections not just for 1-D and 2-D image processing, but also for 3-D reconstruction. Thus, in combination with techniques of image processing, cryoultramicrotomy can give the muscle morphologist the detailed data that are needed to match the molecular biologists, biochemists and immunologists in the interpretation of their data about physiological and pathophysiological events in muscle fibres at the macromolecular level.     

Reconstruct: a free editor for serial section microscopy

Many microscopy studies require reconstruction from serial sections, a method of analysis that is sometimes difficult and time-consuming. When each section is cut, mounted and imaged separately, section images must be montaged and realigned to accurately analyse and visualize the three-dimensional (3D) structure. Reconstruct is a free editor designed to facilitate montaging, alignment, analysis and visualization of serial sections. The methods used by Reconstruct for organizing, transforming and displaying data enable the analysis of series with large numbers of sections and images over a large range of magnifications by making efficient use of computer memory. Alignments can correct for some types of non-linear deformations, including cracks and folds, as often encountered in serial electron microscopy. A large number of different structures can be easily traced and placed together in a single 3D scene that can be animated or saved. As a flexible editor, Reconstruct can reduce the time and resources expended for serial section studies and allows a larger tissue volume to be analysed more quickly.     

Restoration of confocal images for quantitative image analysis

Quantitative studies of three-dimensional (3-D) structure of microscopic objects have been made possible through the introduction of microscopic volume imaging techniques, most notably the confocal fluorescence microscope (CFM). Although the CFM is a true volume imager, its specific imaging properties give rise to distortions in the images and hamper subsequent quantitative analysis. Therefore, it is a prerequisite that confocal images are restored prior to analysis. The distortions can be divided into several categories: attenuation of areas in the image due to self-absorption, bleaching effects, geometrical effects and distortions due to diffraction effects. Of these, absorption and diffraction effects are the most important. This paper describes a method aimed at the correction of diffraction-induced distortions. All the steps necessary in restoring confocal images are discussed, including a novel method to measure instrumental properties on a routine basis. To test the restoration procedure an image of a fluorescent planar object was restored. The results show a considerable improvement in the z-resolution and no ringing artefacts. The relevance of the method for image analysis is demonstrated by a comparison of results of applying 3-D texture analysis to restored and unrestored images of a synthetic object. Furthermore, the method can be successfully applied to noisy fluorescence images of biological objects, such as interphase cell nucei.